N5-MethyltetrahydrofolateHomocysteine Methyltransfmmes ROBERT T. TAYLOR
HERBERT WEISSBACH
. . . . . . . . . . . . . . . . . . . . . . .
. . .
. . . .
I. Introduction . . . . . 11. BE Methyltransferase from Escherichia coli A. Assay and Purification . . B. Physical Properties . , C. Catalytic Properties . . . D. AIkylation Studies with Radioactive N’-Methyl-&-fdate and the Light Stability of a Methyl-Bu Enzyme . E. Studies on the Role of S-Adenosyl-L-methionhe F. Mechanism of N6-Methyltetrahydrofolata-Homocyst.ebe Transmethylation . . . . 111. Non-Bu Methyltransferase from Escherichia coli A. Assay and Purification . . B. Physical Properties . . . . C. Catalytic Properties and Binding of Folate Substrate D. Repression of Enzyme Syntheeis , IV. Other Sources of Non-Bu and & N‘-MethyltetrahydrofolsteHomocysteine Methyltransferma . . . . . A. Non-Bu Methyltransferarres . . . B. Bu Methyltransferns .
.
. .
. .
. . . .
. . . . .
. . . . .
. . . .
. . . . . . . .
. . . . . . . . . . . . . . . .
. . . .
. . . . . . . .
121
122
1M 1W
127 137 143 151 151 161
165 I56
158
160 160
16%
1. Introduction
The terminal reaction in de novo methionine biosynthesis involves a methyl group transfcr from 5-methyl-Hr-folatc (1) to hornocpteine. 1. A bbrevintions : 5-methyl-€I,-f olatc, Z~‘-methyltetrahydrofolate (L-monogluhmate) ; 5-methyl-Hdolate (Glun), l,”’-methyltetrahydrofolate (y-L-trkluhmak) ; 121
122
ROBERT T. TAYLOR AND HERBERT WEISSBACH
In Escherichia coli there are two enzymes with quite different characteristics that catalyze this conversion (la*). The two reactions are shown below :
-
system + homocysteine reducing AMe methionine + H4-folate (Glul, Glur, etc.) Me*+ &Methyl-Hd-folste (Glut, Glur, etc.) + homocysteine phosphate methionine + H,-folate (Glur, Glur, etc.)
S-Methyl-H4-folate (Glul, Glu,, etc.)
-
(1) (2)
Reaction (1), which is catalyzed by a cobalamin-containing protein, also requires catalytic levels of a reducing system and AMe ( 1 ; reviews, references 4 4 ) . Methyl-H4-folates containing one or more glutamates can function as the methyl donor in this reaction. The enzyme that catalyzes reaction (2) does not contain a BIZ (I) prosthetic group, and the only requirement is Mgz+ ions, although a stimulation by inorganic phosphate has been observed. Poly-L-glutamate forms of 5-methyl-H4folate, but not the monoglutamate derivative, can function as substrates in this reaction (6, 7). This report will emphasize the characteristics of both types of methyltransferase isolated from E. coli and the mechanism by which the BIZ-dependent enzyme catalyzes reaction (1). II. BI2 Methyltransfarase from Escherkhia cdi
A. ASSAYAND PURIFICATION 1. Methods of Assav Catalysis of reaction (1) can be determined by measuring either the formation of Hr-folate or methionine. H,-Folate is formed stoichiometriAMe, S-adenosyl-L-methionine ; AH, S-adenosyl-L-homocysteine ; B, is used to denote various cobdamins, e.g., cyano-Bu, cyanocobalamin; methyl-BIZ, methylcobdamin; and prOpyl-Blr, propylcobalamin; &,, a one-electron reduced derivative of aquo-Bu containing Cow; Bus, a two-electron reduced derivative of aquo-Bl, containing Cot+; IGHPOI, potassium phosphate buffer, pH 7.4. la. D. D. Woods, M. A. Foster, and J. R. Guest, in “Transmethylation and Methionbe Biosynthesis” (S. K. Shapiro and F. Schlenk, eds.), p. 138. Univ. of Chicago Prese, Chicago, Illinois, 1965. 2. M. A. Foster, G. Tejerina, J. R. Guest, and D. D. Woods, BJ 92, 476 (1964. 3. J. M. Buchanan, H. L. Elford, R. E. Loughlin, B. M. McDougall, and S. Rosent h d , Ann. N. Y.Acad. S& 112, 766 (lW).
4.
FOLATE METHYLTRANSFERASES
123
cally with methionine (8, 9) and can be converted quantitatively with formic acid to 5,10-methenyl-H4-folate. The latter is measured by its absorption a t 350 nm (9). Measurement of the methionine produced is a more sensitive method of assay. Initially, it was determined by microbiological assay (1&1d) ; however, the chemical synthesis of dl- [5-14C]methyl-H,-folate from dl-H4-folate and [“C] formaldehyde (IS) led to the development of a simple tracer assay in which [l4C]methylrnethionine is separated from the unreacted dl- [ 5-’-’C]methyl-H4-folate by an ion exchange procedure (14). The absolute amount of B,, methyltransferase activity observed depends markedly on the reducing system. But a convenient system for routine assays contains in a total volume of 0.2 ml: KnHP04 ( I ) , 20 pmoles ; dl- [5-14C]methyl-H4-folate (2000 cpm/mpmole) , 30 mpmoles; L-homocysteine, 50 mpmoles ; AMe, 10 mpmoles ; 2-mercaptoethanol, 40 ymoles; cyano-B,2, 10 mpmoles; and enzyme (16).After a 15-min incubation a t 37O, catalysis is terminated’with 0.8 ml of ice cold water. The diluted mixture is placed on a 0.5 X 3.0 cm column of Bio-Rad AG1-XS (chloride) (16,16)and the effluent fluid containing [14C]methylmethionine is assayed for radioactivity. 2. Purification
The BI2protein has been partially purified from extracts of E. coli B grown commercially in cyano-B12 supplemented media (16).A sum4. H. Weissbach and R. Taylor, Fed. Proc., Fed. Amm. SOC.Ezp. BWl. 25, 1649 (1966). 5. H. A. Barker, BJ 105, 1 (1967). 6. H. Weiesbach and R. T. Taylor, Vitam. Horn. (New York) 28, 415 (1970). 7. R. L. Blakley, in “Frontiers of Biology: The Biochemistry of Folk Acid and Related Pteridines” (A. Neuberger and E. L. Tatum, eds.), p. 332. North-Holland Publ., Amsterdam, 1969. 8. A. R. Larrabee, S. Roaenthal, R. E. Cathou, and J. M. Buchanan, JBC 238 1025 (1963). 9. S. Rosenthal, L. C. Smith, and J. M. Buchanan, JBC 240, 836 (1985). 10. B. F. Steel, H. E. Sauberlich, M. 5. Reynolds and C. A. Baumann, JBC 177, 533 (1949). 11. F. Gibson and D. D. Woods, BJ 74, 160 (1960). 12. F. T. Hatch, A. R. Larrabee, R. E. Cathou, and J. M. Buchanan, JBC 236, 1095 (1961). 13. J. C. Keresetesy and K. 0. Donaldson; BBRC 5, 286 (1981). 14. H. Weissbach, A. Peterkofsky, B. G. Redfield, and H. Dickerman, JBC 238, 3318 (1963). 15. R. T. Taylor and H. Weissbach, JBC 242, 1602 (1967). 16. R. T. Taylor, ABB 144, 352 (1971).
124
ROBERT T. TAYLOR AND HERBERT WEISSBACH
TABLE I PURIFICATION OF Blr TRANSMETHYLABE FROM 2
Fraction 1. Extract 2. Manganae chloride
supernatant solution
4. 5. 6. 7.
8. 9.
0
E. mli B*
Volume (ml)
Total activity (k-units)
Protein (mg/ml)
9,400 10,200
10.0 8.9
38.3 20.0
26.5 43.0
100 89
600
4.6
67.0
115
46
3.9 2.9 1.9 1.45 1.1 0'.8
56.0 5.3 10.0 2.4 3.3 1.0
200 555 1,800 4,200 5,500 7,600
3. Ammonium sulfate and
protamine sulfate pH 4.4 precipitation First DEAE-Sephadex Hydroxylep~tih Second DEAE-Sephadex SephadexG200 Third DEAESephadex
KG OF
345 1,ooo 105 145 60 102
Specific activity Yield (units/mg) (%)
39 29 19 14.5 11 8
From Taylor and WeisSbach ( 1 6
mary of this purification is given in Table I. One unit of activity is defined as the amount of enzyme that catalyzes the formation of 1 mpmole of ["Clmethylmethionine per 15 min a t 37" in a 2-mercaptoethanol cyano-B12reducing system (15). The details of the purification and a recent modification of the early step sequences have been published (16, 17). B1,Enzyme that has been subjected to the steps in Table I sediments as a single major species in the ultracentrifuge, yet it is definitely not homogeneous. Cellulose-polyacetate and polyacrylamide-gel electrophoresis further separate these enzyme preparations into several closely spaced protein bands, only one of which is active (16).
+
B. PHYSICAL PROPERTIES 1. Absorption Spectrum and Blz Chromophore
Partially purified preparations of B, methyltransferase (15, 17) contain 1-2.5 mpmoles of firmly bound cobalamin per milligram of protein. Because of this bound B,,, the enzyme prepars tions are distinctly salmoncolored and in the visible region display absorption maxima a t 355, 405, and 475 nm plus a shoulder at 530 nm (15). Their absorption spectra closely mimic that of BIZ, (1, 18). However, unlike BI2,., the enzymebound cobalamin does not yield an electron spin resonance (ESR) spec17. R. T. Taylor, ABB 137, 629 (1970). State Coll. J . Sci. 26, 555 (1952). 18. H. Diehl dnd R. Murie, IOWQ
4.
FOLATE METHYLTRANSFERASES
125
trum, and it is not oxygen labile (16). It has been pointed out (10,&I) that the spcctrum of thc B,? protein also resembles that of a cobalamin a t low pH valucs, where the 5,6-dimethylbenzimidazolylnucleotide is not coordinated to the cobalt (21). Attempts to identify the cobalamin chromophore by treatment with hot ethanol in the dark have yielded sulfito-B,, as the major corrinoid in the extracts (19, 22). It remains questionable, though, whether the enzyme as usually prepared (16, 17) actually does contain sulfito-B,Z or whether the sulfito-B,, is an artifact of the alcohol stripping and subsequent identification procedures (19, 29).Sulfito-Blz is a light-labile cobalamin which forms nonenzymically by exposing aquo-Blz to bisulfite (23).The salmon-colored B,, protein will react with alkaline cyanide to give a completely different absorption spectrum that ig typical 'for the dicyano derivative of vitamin B,, (16). Subsequent extraction of the chromophore from the protein then yields only dicyano-B,, (19). 2. ResolutiowReconstitution and Molecular Weight
Selective treatment of the B,, protein with 6 M urea + l,4-dithiothreitol resolves it into a colorless apoprotein plus Blzr (17). Release of the bound cobalamin as BIZ,, rather than sulfito-B1,, results from its reduction by the 1,Cdithiothreitol which is added to stabilize the much more labile apoenzyme ( 9 3 ~ Upon ). Sephadex G-25 filtration, apoenzyme retaining only about 5% of the original B,, is separated from the resolution mixture (17).Incubation of the resolved apoenzyme with methyl-BIZ results in the spontaneous formation of a methyl-BIz holoenzyme. The binding of methyl-B,, resulting in the transformation of apoenzyme into holoenzyme requires no ancillary protein fractions or cofactom, and it is temperature dependent (17).From Table I1 it can be seen that reconstitution requires a complete corrinoid, i.e., a cobalamin. The ligand a t the sixth coordinating site is equally important since methyl-B,, yields predominantly holoenzyme but deoxyadenosyl-B12and cyano-Blz promote no conversion. Sulfito-BIZis bound, but its sulfonate group is reductively cleaved when one tests for holoenzyme activity in a thiol-containing reducing system (17).From these resolution-reconstitution studies it was concluded that all the bonds between the B,, and the apoprotein must be noncovalent in nature. 19. R. Ertel, N. Brot, R. Taylor, and H. Weiasbach, ABB 126, 353 (1968). 20. H. P. C. Hogenkamp, Annu. Rev. Biochem. 37, a25 (1988). 21. J. A. Hill, J. M. Pratt, and R. J. P. Williams, J . Theor. Biol. 3, 423 (1962). 22. S. Takeyama and J. M. Buchanan, J . Bbchem. (Tokyo) 49, 578 (1961). 23. F. Wagner, Annu. Rev. Biochem. 35, 405 (1966). 238. Various forms of the Ba protein are defined aa in reference 17.
126
ROBERT T. TAYLOR AND HERBERT WEISSBACH
TABLE I1 CONVERSION OF UREA-RESOLVED APOENZYMETO HOLOENZYME WITH VARIOUS CORRINOIDS~~~ Compound present in preliminary incubation None Methyl-Bit (or propyl-Bl9 and subsequent photolysis) Deoxyadenosyl-Bld SUlfito-Bis Photolyzed ~UlfitO-Bi~ Hydroxy-B19,10-70 mpmoles Cyano-Bii Diaquocobinamide Methylcobinamide, 10-50 mpmoles a
Holoenzyme
(%I
3 77-80 4 48-52 15-18 8-13 6 11 4 3-4
From Taylor (17).
* Mixtures (0.2 ml) containing apoenzyme, 1.0 mg; the indicated corrinoid compounds,
10 mpmoles of each except where noted; and KnHP04,pH 7.4,20 pmolea, were incubated in the dark 5 min at 37". Unbound corrinoids were removed with Norite cellulose columns, and the percentages of holoenzyme were assayed (17,$4) with identical aliquots of the column effluent solutions, both before and after they were illuminated (100 W, 15 min, 20 cm,0"). 0 Bib waa generated by photolysis of a duplicate deoxyadenosyl-Bn system under an H, atmosphere and was incubated under HI gas.
In vitro, the conversion of initial BIZ holoenzyme (2%) into apoenzyme and back to a methyl-B,, holoenzyme is accompanied by a gross, but reversible, change in the shape of the protein (17,25). Initial (purified) B,, holoenzyme has a sedimentation coefficient of 7.0s (16, 17). Apoenzyme has a sedimentation coefficient of only 6.2 S (17) but a larger Stokes radius than either the initial or the reconstituted forms (23~)of holoenzyme (Table 111).Sephadex G-200 chromatography of apoenzyme in 6 M urea + 1,4-dithiothreitol gives no indication that loss of the BIZchromophore causes the apoenzyme to dissociate into subunits (%). Based on their sedimentation coefficients in sucrose gradients and their elution from a calibrated Sephadex G-200column, all forms of BIZ holoenzyme have a molecular weight of 140,000-150,000 (15,17, 65). A 9-A decrease in the Stokes radius of apoenzyme (Table 111) cannot be effected with methylcobinamide (25) which also cannot transform apoenzyme into active holoenzyme (Table 11).An attachment between the 5,6-dimethylbenzimidazolyl nucleotide and the apoprotein is appar24. H. Weieabach, B. Redfield, and H. Dickerman, BBRC 17, 17 (1964). 25. R. T. Taylor, BBA 242, 356 (1971).
4.
127
FOLATE METHYLTRANSFEBASES
TABLE 111 STOKESRADII OF UREA-RESOLVED APOENZYME AND THREE HOLOENSYMES DETERMINED BY SEPHADEX (3-200 CHROMATOORAPHY~
Form of enzyme
No. of experiments
Initial B I holoenzyme ~ Urea-resolved apoenzyme Reconstituted methyl-Blp*H holoenzyme Propyl-BII enzyme
Stakea radid
KD
(A)
flfo
6 4 4
0.215 0.156 0.219
54.2 f 1.1 62.9 f 0.3 53.6 f 0.4
1.55 1.79 1.53
4
0.209
55.1 f 0.3
1.67
~~
~
~~~
~
From Taylor ($6). are the mean f S.D. c Prepared by cobalt akylation with propyl iodide and 90% propylatd (H). It was chromatographed in the dark and then column fractions were asgnyed after photolpia of the cobalt-propyl bond (97). 0
b Valuea
ently necessary in order to return the loosely folded apoenzyme to a more compact structure. Using a molecular weight of 14O,OOO, purified emyme with 2.5 ,mpmoles/mg of bound B,, (Table I, fraction 9) would contain 0.35 mole of cobalamin per mole of protein. The estimated purity of such a preparation is, therefore, only about 3376, assuming that there i s only one molecule of bound B,, per molecule of enzyme. It was reported by Stavrianopoulos and Jaenicke ($6) that a homogeneous preparation of B, enzyme was obtained from E. coli (strain unspecified) by a series of steps similar to those in Table I. Its molecular weight, too, was about. 14O,O0OJ although its B,, content of 0.5 mole/mole of protein ($6) would not be compatible with a homogeneous protein.
C. CATALYTIC PROPERTIES 1. Specific Inhibition b y Propyl Iodide Although the cobalamin chromophore is tightly bound to the apoprotein, its cobalt atom still possesses the chemical reactivity (93) of unbound aquo-BI, or Cyano-B,,. In particular, the cobalt can be reductively alkylated with propyl iodide ($7, 98) to give an inhibited, propyl-B1, enzyme [reaction (3)]. 26. J. Stavrianopoulos and L. Jaenicke, Eur. J. Bwohem.3, 95 (1967). 27. R. T.’Taylor and H. Wekbach, JBC 842, 1509 (1967). 2.8. N. Brot and H. Weissbach, JBC 240, 3064 (1985).
128
ROBERT T. TAYLOR AND HERBERT WEISSBACH
I
>?: Enz
+
CH,-CH,-CH,-I
Initial B,, enzyme
reducing eyatem dark
CH,-CH,-CH,
~
\I/
'";". Elu
+ I
(3)
Inactive pmpyl-B, enzyme
Reversal of the inhibition by a short exposure to light (28) provided a strong clue that the site of blockage in unpurified enzyme preparations is the bound B,, group. Alkyl-BIZ compounds are extremely photosensitive because of the ease with which a carbon-to-cobalt bond is cleaved by light under aerobic conditions (63). When millimicromole quantities of purified B,, enzyme became available (16), direct proof was obtained (67, 6s) for reaction ( 3 ) as well as reaction (4).
Inactive pmpyl-B, enzyme
Active aquo-B,, enzyme
Treatment of purified enzyme with propyl iodide in a FMNH, + 1,4dithiothreitol reducing system yielded an inactive B,, protein with an altered chromophore spectrum. Light restored activity and simultaneously generated an absorption spectrum resembling that of aquo-B,? (67). When [ l-14C]propylbromide was the alkylating agent, the inhibited B,, protein was radioactive, but most of the 14C was reversibly lost upon photolysis (67). By extracting the cobalamin from propyl iodide-treated enzyme in the dark, it was possible to identify the prosthetic group by spectral and paper chromatographic procedures as propyl-B,, (69). The ability to inhibit selectively the B12 methyltransferase with propyl iodide and reverse the inhibition with light has been a powerful aid in establishing whether or not this enzyme is involved in the catalysis of other reactions. If the reader refers to either of the two initial papers (67,68) on chemical propylation of E. coli B,, methyltransferase, an oLvious discrepancy will be noted regarding the incubation conditions needed. S-Adenosyl-Lmethionine was required in the earlier study using unpurified BIZenzyme (68),but it actually prevented propylation of the purified B12protein in a FMNH, 1,4-dithiothreitol reducing system (67). This anomoly has been ascribed to the presence of contaminating enzymes in the unpurified
+
29. R. T. Taylor and H.Weiasbach, ABB 123, 109 (1968).
4.
129
FOLATE METHYLTRANSFERASES
Blz enzyme preparation (30)which degraded AMe to homocysteine. It
will become apparent latcr in the discussion why the inhibition of propylation by AMe ($7, So) provided an important clue as to the role of AMe in reaction (1). 2. Methyl Group Transfer Reactions Catalyzed by the Enzyme
The biosynthesis of methionine via reaction (1) is the important catalytic function of the E. coli BIZmethyltransferase. Table IV illustrates its unique dependencies.on both a reducing system and AMe. While 2-merTABLE IV REQUIRE~~ENTS FOR METHYLGROUP"MNSFER FROM [5-"C]METHYG&-FOLATE TO HOMOCYBTEINE~ Methionhe formed (mrmol4
Reaction mixture ~
~~
~
Experiment A* Complete system (HI) -Enzyme -ZMercaptoethanol -CyanO-BII -AMe -Homocysteine Complete system (aerobic) Experiment Bd Complete system (Hl) -Enzyme -1,4dithiothreitol -AMe -Homocysteine -FMNHnand platinum
6.4 0
0.08
1.2
0.08 0.8~
4.4 4.7 0
2.9 0.02 0.14 0.14
From Taylor and Weissbach (16). The reaction conditions for experiment A are described in the text in Section IIJA,lJ except that an H2 gas phase was used unlm indicated otherwise. Mixtures (0.2 ml) ~ and were incubated for 15 min at 37". contained 2.4 pg of B I enzyme c In the absence of homocysteine, 2-mercaptoethanol(O.2 M )functions ale0 as a methyl group acceptor. In the complete system, however, saturation of the B I enzyme ~ with homocysteine (0.25 mM) completely eliminates any methyl group transfer to Zmercaptoethanol and renders this w a y method specific for radioactive methionine. "or experiment B the complete system (0.2 ml) contained the same amounta of buffer, AMe, homocysteine, and dL[5-~4C]methyl-&folate as in experiment A, but of Blr enzyme. The reducing system consisted of FMNHt, 50 mpmoles; only 0.7 platinum, 0.1 mg; and 1,4dithiothreitolJ 5.0 pmolea under a Hsgas pheae. a
b
30. R. T. Taylor, C. Whitfield, and
H.Wekbach, ABB 125,
240 (1988).
130
ROBERT T. TAYLOR AND HERBERT WEISSBACH
captoethanol f cyano-B,? can be used as a convenient artificial reducing system ( l a ) , FMNHz 4- 1,Cdithiothreitol (Table IV) has consistently been observed (15, 51-SS) to sustain the highest rate of catalysis. Most investigators have used FADH,, generated either enzymically (26, S4, S6) or chemically (S,S6, S6). However, in the presence of 1,4-dithiothreitol and NADH, reduced flavin bound to diaphorases and lipoamide dehydrogenases will substantially replace the need for free FMNH, or FADH? (S3).The coproduct of reaction (1) (H4-folate) will also serve, a t high concentrations, one-third as well as FMNH, as a reducing agent (31). Both AMe and reduced flavin function catalytically in reaction (1) (14, 34, S6),the latter even when it is bound in a flavoprotein (33).When one uses a 2-mercaptoethanol + cyano-BI2 reducing system (Table IV), the vitamin only functions nonenzymically to accelerate reduction by the thiol (3)as described by Peel (87). Galivan and Huennekens (38) reported that a 3000 molecular weight “S” protein derived from extracts of E. coli K-12is required for the Blz protein to catalyze reaction (1). They concluded that the “S” protein is an essential subunit of the latter, but exogenous methyl-B,, (or aquo-Blz) plus a dithiol compound or an NADH-dependent flavoprotein fraction were also essential in their system (S8).It seems probable that the “S” protein is merely a component of the supplementary reducing system since no evidence for an essential “S” protein has been found by other investigators (9,16,96,34,S9) who have used purified E. coli B,, enzyme. In the presence of an excess of Bl2 enzyme, reaction (1) is catalyzed essentially to completion ( 8 ) ,apd for all practical purposes it is unidirectional. Rudiger and Jaenicke (40) found it to be slightly reversible and published an equilibrium constant of 1.4 X lo5 in the forward direction. 5-Methyl-H4-folate containing one or more L-glutamates can servc as a substrate in reaction (1) (8); however, Fi-methyl-&-folate (Glu,) 31. R. T. Taylor and H. Weissbach, ABB 129, 745 (1969). 32. R. T. Taylor and M. L. Hanna, ABB 137, 463 (1970). 33. R. T. Taylor and M. L. Hanna, ABB 139, 149 (1970). 34. S. Rmnthal and J. M. Buchanan, Acta Chem. Scnnd. 17, Suppl. 1,288 (1963). 35. M. A. Foster, M. J. Dilworth, and D. D. Woods, Nature (London) 201, 39 (1964). 36. H. L. Elford, H. M. Katzen, 13. Rosenthal, L. C. Smith, and J. M. Buchanan, in “Transmethylation and Methionhe Biosynthesis” (S. K. Shapiro and F. Schbnk, eds.), p. 157. Univ. of Chicago P r e ~ Chicago, , Illinois, 1966. 37. J. L. Peel, JBC 237, PC263 (1982). 38. J. Galivan and F. M. Huennekens, BBRC 38, 46 (1970). 39. L. Jaenicke and H. Riidiger, Fed. Proc., Fed. Amer. SOC.Ezp. Bwl. 30, 160 (1971). 40. H. Riidiger and L. Jaenicke, FEBS Lett. 4, 316 (1969).
4.
131
FOLATE METHYLTRANSFERASES
was reported to be only one-half as active as 5-methyl-H4-folate (Glui) (96‘). Chemically synthesized dZ-5-methyl-Ha-folate (13, 41) contains an asymmetric center a t carbon 6 in the pteridine ring. It is introduced when folic acid is chemically reduced to H4-folate (M). BIZMethyltransferases distinguish between these isomers and stereospecifically remove methyl groups from only l-5-methyl-H4-folate (8, 43, 4.6). It must be stressed that designating the active isomer as I (8, 43, 4.4) does not refer to the direction that it rotates light. It is merely based on the fact that the active isomer of 5-methyl-H4-folate is derived ensymically from E-HI-folate which is ( - ) or levorotary. Despite a recent paper (46) deducing that d- and Z-5-methyl-H4-folate rotate light in different directions, direct 12.9” and measurements have, in fact, yielded specific rotations of +4.0” for the 2 (active) and d (inactive) isomers, respectively (46). Under saturating conditions with respect to the other components the K , values for 1- [5-14C]methyl-a-folate, homocysteine, and AMe were 30, 16, and 1.6 (47).These K , values were obtained when purified enzyme was used to catalyze reaction (1) aerobically in a Z-mercaptoethanol cyano-BI2 reducing system. In a FMNHz 1,Cdithiothreitol reducing system (16))the K , for l-[l%]methyl-H,-folate, 35 does not change significantly (&), but the apparent K , for AMe decreased about 10-!&fold (32). Rosenthal et a2. (9) first reported that 2-mercaptoethanol wouId substitute, though poorly, for homocysteine as a folate methyl group acceptor [reaction ( 5 ) 3.
+
+
+
bMethyl-H,-folate
a,
-
+ Zmercaptoethanol reduciru AM0
system
S-rnethylmercaptoethanol+ Hcfolste (5)
At an optimal concentration of 0.1 M 2-mercaptoethanol the rate of transmethylation was 8-fold slower than when a saturating level of homocysteine (0.25 mM) was present (9, 16). At equal concentrations of 10 mM, 2-mercaptoethylamine, thioglycolic acid, cysteine, glutathione, and 4-mercaptobutyric acid all displayed less than 10% of the acceptor 41. W.Sakami and I. &tins, JBC e3s, PCW (lW1). 42. C.K. Mathews and F. M, Huennekens, JBC 83!5, 3304 (1960). 43. K. 0.Donaldeon and J. C. Keresstesy, JBC 837, 3816 (1962). 44. B. T. Kaufman, K. 0. Donaldson, and J. C. Kereategy, JBC 238, 1498 0963). 45. H. Riidiger, FEBS Lett. 11, 286 (1870). 46. J. C. Keresztesy and K. 0. Donaldson, Iowa State CoU. J . Isci. 38, 41 (1963). 47. R.T. Taylor and H. Wehbach, “Methods in Ensymology,” Vol. 17B, p. 379, 1871. 48. R. T. Taylor and M. L. Hanna, ABB 151, 401 (1Sm).
132
ROBERT T. TAYLOR AND HERBERT WEISSBACH
activity of 2-mercaptoethanol (9).Dithiols such as 2,3-dimercaptoethanol (9) and 1,Cdithiothreitol (Table IV, experiment B) have negligible methyl acceptor activity. In addition to reactions (1) and (5), preparations of E. coli B,, enzyme also catalyze reactions (6), (7), and ( 8 ) . AMe
+ homocystthe FMNHI + 1,4-dithiothreitol+ methionine + AH
-
+ Hd-folate FMNAI + 1,rl-dithiOthreitOl5-methyl-Hd-folata + AH aerobic Methyl-BIs + homocyshine methionine + aquo-Blt dark
AMe
t
(8)
Reaction (6)was also first detected by Rosenthal et al. (9) who demonstrated that the coproduct was AH ( 1 ) . Catalysis requires a reduced flavin and like reaction (1) (Table IV) it is stimulated by 1,Pdithiothreitol (16).Even in the presence of saturating levels of both substrates, it is a very slow enzymic reaction. Reaction (7) was first observed qualitatively by Stavrianopoulos and Jaenicke (26). It was subsequently studied in detail by Taylor and Weissbach (31, 49). The folate reaction product is the active 1 isomer of 5-methyl-Hc-folate (49) and the K , for dl-H,-folate is 0.17 mM (31). Table V summarizes the relative rates (in terms of the bound B12)and the apparent K , values of AMe for reactions (1), (6), and (7), respectively. Strikingly, 5-methyl-H4-folate is the preferred methyl donor in spite of the low K , of AMe both as a cofactor [reaction (1) and as a substrate [reactions (6) and (7)]. Guest et al. (60) first observed the catalysis of reaction ( 8 ) . It occurs in the dark under both aerobic and anerobic conditions and requires no cofactors (14,15,50).Under anaerobic conditions, the cobalamin product that accumulates stoichiometrically to the methionine is B,,, (51-53). Under aerobic conditions aquo-B1, accumulates (51) as a result of the oxidation of BIZ,.It is not possible to designate BIzr as the initial cobalamin product of reaction (8) because it may be formed by the anaerobic, spontaneous reaction of aquo-B,, with the homocysteine substrate (60, 6 1 ) . Alternatively, BIZ,may arise as a result of the rapid oxidation of B12, ( I ) to BIZ,.or the reaction of B12,with water (54).Reac49. R. T. Taylor and H. Weissbsch, ABB 129, 728 (1969).
60. J. R. Guest, 5. Friedman, D. D. Woods, and E. L. Smith, Nature (London) 195, 340 (1962). 51. H. Weissbach, B. G. Redfield, and H. Dickerman, JBC 239, 1942 (1964). 62. 8. S. Kerwar, J. H. Mangum, K. G. Scrimgeour, J. D. Brodie, and F. M. Huennekens, ABB 116, 305 (1966). 53. J. R. Guest, S. Friedman, M. J. Dilmorth, and D. D. Woods, Ann. N . Y . Acad. Sci. 112, 774 (1964). 54. 8. L. Tackett, J. W. Collat, and J. C. Abbott, Biochemistry 2, 919 (1963).
APPARENTK , Reaction
+ + + + + +
(1) [5J*C]Methyl-&folate homocysteine -+ [Wlmethylmethionine H4-fOhh (6) p4C]Methyl-AMe homocysteine + [Wlmethylmethionine AH (7) [WlMethyl-AMe Hcfolate + [5J4C]methyl-&fohte AH
(OR
x
s
TABLE V 50% REACTION CONCENTRATION) FOR AMP
c3
?2
Ba enzyme concn.
[BoundBlz]
Turnover numbee
5 . 2 X 10-* M
3.0 X 103 M
17
500-780
3.9
5 . 0 X 1W8M
2.6 X lW'M
19
2.2
15
21
1.3
31
Apparent K,
3.7
+
x 1 0 - 8 ~ 1.8 x
10-7~
K m
Ref.
a Each appamnt K, was determined in the FMNH@t) 1,44thiothreitol reducing system cited in the legend to Table IV and described in detail in reference 16. All reaction mixtures (0.2 ml) contained 20 moles of K,Hpo4, pH 7.4, and all incubations were for 15 min at 37" with continuous Hs Bseeing (16). Millimicmmoles of p4C]methyl transferred/mh/mpmole of bound BIZin the presence of a saturating concentration of A M e (or [Wl-
*
methyl-AMe), 50 pM.
134
ROBERT T. 'TAYLOR AND HERBERT WEISSBACR
tion (8) can be balanced if one depicts the immediate cobalamin product as B,,, [reaction (9)]
-
+ homocysteine dark methionine + Bls. + H+
Methyl-B1*
(9)
as has been suggested (7, 16,61,69). I n view of the fact that a completely homogeneous B,, protein preparation has not been obtained, one might question whether reactions (1) , and (5)-(8) are indeed catalyzed by the same enzyme. Over a 7-fold purification of enzyme from E. coli 113-3, a constant ratio was observed between the activities for reactions (1) and ( 5 ) (9). Likewise, for over a 150-300-fold enrichment of enzyme from E. coli B, constant ratios were found between the activities for reactions ( l ) ,(6), and (8) ( 1 6 ) . While these findings were strongly suggestive of a single enzyme, the most compelling evidence was provided by chemical propylation data. Table VI shows that, except for reaction (8), all of these reactions were inhibited in a light-reversible manner and to the same extent by propyl iodide. Since propyl iodide specifically alkylates the B,,-cobalt (!27), one must conclude that the cobalamin enzyme is the catalyst for a t least reactions (1) and (5)-(7). The insensitivity of reaction (8) to propylation (97,66),despite the TABLE VI EFFECTOF PROPYLATION ON FIVEREACTIONS CATALYZED BY A PURIFIED PREPARATION OF E. cbli B METHYLTRANSFERASE~
[WIMethyl group transfer reaction
+ + +
(1) [5-W]Methyl-H~-folate
homocysteie 4 [W]methylmethionine Hcfolate ( 5 ) [5W]Methyl-H4-folate Zmercaptoethanol+ 5-[~~C]methyl-mercaptoethanol+ H4-folate (6) [14C]Methyl-AMe homocysteine + [W]methylmethionine AH (7) [W]Methyl-AMe Hcfolate + [5-W]methyl-&-folate AH ( 8 ) [W]Methyl-B~* homocysteine -+ [14Cbethylmethionine aquo-B~~.
+ + +
+ + +
Propylation or inhibition in the dark (%) 92 86 87 90
0
a Separate samplea of Ba enzyme were incubated in the dark with propyl iodide in a FMNH, (Pt) 1,4-dithiothreitol reducing system (fl).Appropriate aliquot8 from these mixtures were then assayed (16, 31) for the indicated transmethylase activities before and after being exposed to light (87). For each reaction, the percentage of propylstion was eatimated from the ratio of activity given by an aliquot of enzyme in the dark relative to an identical aliquot which had been photolyzed (87).
+
4. FOLATE METHYLTRANSFERASES
135
copurification of this activity (16), suggested that two distinct sites on the B,, protein might catalyze reactions (1) [plus (5)-(7)] and (8), respectively. Recently, additional evidence for this interpretation was obtained. A 200-fold purified preparation of the B,, enzyme was further fractionated to give a small amount of material which appeared to be predominantly a single protein upon disc gel electrophoresis (16). This protein retained the ability to catalyze both reactions (1) and (8) with no change in the ratio of the activities. Moreover, a propyl-BlS enzyme was shown to catalyze reaction (8) with no loss of the propyl group. Also, reconstituted holoenzyme containing tritium in the cobalamin group catalyzed this reaction with little loss of its radioactive chromophore (16). The apparent K,,, of apoenzyme for methyl-Bln as a prosthetic group [reaction ( l ) ]is 2.0 while the K,,, of holoenzyme for methyl-Bln as a substrate is >2.5 mM. These results and other kinetic data (16) indicate that [in the catalysis of reaction (8)] methyl-BlZ and homocysteine form a ternary complex with a site on the enzyme separate from where the tightly bound B12 is attached. After the ternary complex has reacted to yield methionine and B12, (or its oxidation product B12,), this loosely bound cobalamin product must be replaced by a new molecule of methyl-B12before catalysis [reaction (8) ] can continue. The site for reaction (8) can apparently use exogenous methyl-BIZ as a substrate because of its lower affinity for cobalamins compared to the site for reaction (1).In contrast to the reaction (1) site, the cobalamin site for reaction (8) probably lacks specific attachment points for the 5,6-dimethylbenzimidazolylnucleotide. Consequently, methylcobinamide is reported to function equally as well as methyl-BIZ as a substrate in reaction (8) (36,65). It will be recalled from Table 11, however, that methylcobinamide will not bind tightly with apoenzyme a t the reaction (1) site and form’holoenzyme. Irrespective, the cobalamin site for reaction (8) does have in common with the site for reaction (1) a high degree of specificity for transferring cobalt-methyl groups. Neither ethyl-B12 (65,66)nor propyl-B,, (16) can function as alkyl donors to homocysteine in reaction (8). Similarly, a propyl-B12 enzyme is inhibited with respect to reaction (1) because the cobalt-propyl group is totally unreactive with homocysteine (16, 17).Reaction (8) is catalyzed over a broad pH range (6-10) with optima at pH 7.5 and 9.5 in phosphate and carbonate buffer, respectively (16). It is an irreversible reaction (61,63).Reaction (1) is catalyzed optimally only near pH 7 (9, 16).
a,
55. N. Brot, R. T. Taylor, and H. Weissbach, ABB 114 258 (1966). 56. H. Weissbach, in “Transmethylation and Methionine Bioaynthesis” (8.K. Shapiro and F. Schlenk, eds.), p. 179. Univ. of Chicago Preess, Chicago, nfinok, 1965.
136
ROBERT T. TAYLOR AND HERBERT WEISSBACH
3. Binding of Radkactive Folate Substrate When purified B,, protein is incubated with dl- [ 5-14C]methyl-H,-folate in the absence of both AMe and reduced flavin, an enzyme-14Ccomplex is formed (48).While the amount of complex formed is independent of the temperature between 0" and 37", its subsequent recovery is quite temperature-dependent. Sephadex G-25 filtration a t 22" yields virtually no complex, but filtration a t 2O-4" results in the isolation of 0.1-0.15 mpmoles of bound "C per millimicromole of bound BI2 (Fig. 1 ) . The low yield of complex resulted from its reversible dissociation upon passage through the column. Since the same amount of complex was obtained with dl-5-methyl-H,- [2-14C]folate and the 14C stripped from heat-denatured enzyme chromatographed like authentic dl-5-methyl-Ha-folate, it was concluded that the folate substrate can bind intact to the B,, protein. Moreover, only the active I isomer of [5 - W ] methyl-H,-folate was bound to the enzyme. By using substrate equilibrated G-25 columns, the maximum amount of binding observed was 0.81 mpmole/mpmole of bound BE. From a study of the equilibrium binding versus the folate substrate concentration, a dissociation constant of 6.2 was calculated, and the
1 -
B X
g
\
2;
% .
1
16.0 (A 1
-
1
-
7
-
1
14.0
12.0 10.0 0.0
'5
df"
6.0 4.0
$
2.0
F
C
E u
m
10
15
20
25
39
35
40
45
50
5
1 0
15
20
Volume, rnl
Wa. 1. Sephadex G-25 filtration of preformed initial Blt CUClenzyrne complex. (A), Systems (0.2 ml) containing &HP04, 20 pmoles; initial Blz enzyme, 7.7 mpmoles; and dZ-[5-1'Clmethyl-~-folate (34,000 cpmlmpmole), 10 v o l e s , were incubated 5 min at 0". KHPO, buffer, 0.8 ml of 10 mM, was added and the entire mixture waa chromatographed a t 2'4". Aliquots were counted for radioactivity and assayed for transmethylase activity. Enzyme was omitted for the control filtration. (B), Procedure aa in (A) except that the preliminary incubations were 5 min at 37', and one filtration was made a t 2 2 O rather than 2'4'. Only the C ' elution profile was determined. From Taylor and Hanna (48).
4.
137
FOLATE METHYLTRANSFERASES
theoretical number of binding sites was 0.92 per molecule of bound B,, (48).Folate has a much lower affinity for the B1, protein than Z-5-methylH4-folate. Interestingly, urea-resolved apoenzyme possessed 90% as much [5-"C] methyl-H,-folate binding activity as the original Blz holoenzyme from which it was prepared (48).It therefore appears that the folate substrate attaches initially to a special protein site rather than directly to the BIZchromophore. At all pH values where the B1, enzyme catalyzes reaction (1), 5-methyl-H,-folate (isoelectric pH = 3.89 & 0.04 S.D.) binds with its W-amino group in its free base rather than its protonated form (48).
D. ALKYLATIONSTUDIES WITH RADIOACTIVE 5-METHYL-&-M)LATE AND THE
LIGHTSTABILITY OF
A
METHYL-B~,,ENZYME
Since the B,, enzyme catalyzed methyl transfer from methyl-B,,, it naturally followed that a methyl-B,, enzyme could be an intermediate in 5-methyl-Ha-folate transmethylation [reaction (1)1. The fact that propylation of the B,,-cobalt blocked reaction (1) also suggested this possibility, but direct evidence was needed. [W]Methyl-H4-folate was, therefore, incubated with stoichiometric amounts of Blg enzyme in a FMNH, + 1,Pdithiothreitol reducing system containing AMe (Table VII) . The amount of radioactive enzyme was determined by precipitating the protein with cold 10% trichloroacetic acid i.n the dark (67).l C Labeled precipitates were then collected on a Millipore filter for liquid scintillation counting. Increasing the concentration of [5-14C]rneth~1-H~folate to 30 & (not shown) gave maximal precipitable counts, namely, 0.5 mpmole of 14C per millimicromole of B12 protein (68).As seen in Table VII, the use of acid precipitation and filtration in the dark (67)removes both the unreacted folate substrate as well as any that is reversibly bound intact (Fig. 1). Ehperiments using variously labeled 5-methyl-&-folate showed that only the methyl-labeled substrate reacted with the enzyme to form an acid-stable radioactive species (Table VIII) Although the results in Tables VII and VIII prove that the folate methyl group can be transferred to an acid-stable position in the presence of the cofactors, these findings did not establish either the location of the methyl group on the enzyme or its reactivity with homocysteine. For these purposes, enzyme was first methylated with [5-14C]methyl-H,-folate and then separated,
.
57. R. T. Taylor and H. Weissbach, ABB 110, 572 (1967). 58. R. T. Taylor and H. Wehbach, ABB 193, 109 (1968).
138
ROBERT T. TAYLOR AND HERBERT WEISSBACH
TABLE VII FOR THE FORMATION OF TRICHMROACETIC ACID-PRECIPITABLE REQUIREMENTS ["C]BltTRANSMETHYLASE WITH [!k"C]METEYGHd-FOLATE5
Reaction mixture
Trichloroacetic acid precipitate (CPm)
Complete system, 0% Complete system, 37" -Enzyme -FMNHa and platinum -1,4dithiothreitol -FMNHa, platinum, and 1,4dithiothreitol -AMe +Homocysteine
187 6219 147 1278 2369 147 276 180
14C bound (mpmole) <0.01 0.51 0.09 0.18
-
0.01 <0.01
From Taylor and Weissbach (67). Complete system (0.2 ml) contained KzHPO4, pH 7.4, 20 pmoles; dl-[5-*4C]methylHd-folate (12,300 cpm/mpmole), 3 mpmoles; AMe, 10 mpmoles; li4-dithiothreitol, 5 pmoles; FMNH?, 50 mpmoles; platinum, 0.1 mg; and B I transmethylase, ~ 0.6 mg. Homocysteine, 0.1 pmole, was added to t,he complete system as indicated. 'Incubations were for 10 min at 37" under eont,inuousHn gassing. c The amount of Blnenzyme added cont,ained about 1.4 mpmoles of bound B1t chromophore. a
b
without denaturation, by passage through a Sephadex G-25 column at 22" (Fig. 2). A large peak of radioactivity is associated with the enzyme (peak I), and the reaction is dependent on AMe. Peak I1 in the figure is an oxidised degradation product of the folate substrate, and peak I11 is unreacted dl- [5-14C]methyl-Ha-folate. Methylated enzyme isolated as TABLE VIII FORMATION OF TRICEIMROACETIC ACID-PRECIPITABLECOUNTSPER MINUTE WITH THREE D I F F E ~ N T LLABELED Y &METHYGH~-FOLATES".L
Folate in complete system
Trichloroacetic acid precipitate (Cpm)
(1) [5-W]Methyl-€€-folate 16,500 cpm/mpmole -Enzyme (2) 5-Methyl-H4-[2J4C]folate 9300 cpm/mpmole -Enzyme (3) 5-Methyl-*Hh-folate 22,000 cpm/mpmole -Enzyme
6222 128 374 352 235 179
~~
~~~~
~
Isotope bound (mpmole) 0.37 0
<0.01
0 <0.01 0
~
From Taylor and Weissbach (68). conditions were the same as in Table VII except that all the systems contained 10 mpmoles of labeled dZ-5-methyl-Hd-folate and 0.52 mg of enzyme. a
b Incubation
4.
139
FOLATE METHYLTRANSFERASES
*O-O
t
Volume (mi)
FIQ.2. Isolation of trichloroacetic acid-precipitable C1'CIB~~-transmethyl~ by Sephadex filtration. A reaction mixture (028 ml) containing 2.0 mg of & enayme, 6 pmoles of 1,44thiothreitol, 10 nytmoles of dGIB'~lmethyl-E-folate ( 1 3 e cpm/mpmole), and all the other components of the complete Wtem in Table V I I were incubated 10 min at 37" under E.Then 0 8 ml of water was added, and the entire mixture was applied to 8 Sephadex G-25 column (for details, see reference 67). A second identical reaction mixture minus AMe waa treated in the same manner.
in Fig. 2 readily transferred most of its [14C]methyl to homocysteine without any further requirement for AMe or a reducing system (Table IX). Transfer was not impaired by either oxygen or light and is actually a very rapid reaction. The 15-min incubation used initially in Table IX is not essential because transfer is largely complete in <5 sec (31). Other experiments firmly established that the homocysteine reactive enzyme (Table IX) contained a ["C]methyl-BIZchromophore. Substrate methylation shifted the major absorption peak in the visible portion of the spectrum from 475 to 520 nm (Fig. 3). This new maximum plus the double peaks at 340 and 380 nm are characteristic of methyl-BlZ (69) whose structure is shown in Fig. 4. Addition of homocysteine essentially regenerated the spectrum of the original Blz ensyme prior to methylation 59. 0. Muller and G. Miiller, Bwchem.
2. 336, a99 (1W2).
140
ROBERT T. TAYLOR AND HERBERT WEISSBACH
TABLE IX REOUIREMENT FOR [l'C]METHYLMETHIONINE FORMATION FROM [14C]B1+TRAN8METHYLAsE"
Experiment
Reaction mixturg
(wd
Methionine formed (mwole)
1. Dark (anaerobic)
Complete system ([W]epzyme homocysteine) -Homocpteine Complete system AMe reducing system
2000
0.16
160 1900
0.01 0.15
Complete system ([Wlenzyme homocysteine) -Homocysteine
2650
0.21
230
0.02
+
+
2. Light (aerobic)
+
~~
a
+
~
From Taylor and Weissbach (67).
* Complete systems (0.3 ml) contained [W]enzyme, isolated as described in Fig. 2,
2600 cpm (Exp. 1) or 2730 cpm (Exp. 2) and homocysteine, 0.2 mole. In Exp. 1, AMe,
lOrmolea, and a reducing system were included in the reaction mixture as indicated. The reducing system consisted of FMNHs, 50 m e o l e s ; platinum, 0.1 mg; and 1,4-dithiothreitol, 5 pmoles. Incubations were for 15 min at 37" under HI gas in the dark (Exp. 1) or aerobic for 15 min at 37" in room light (Exp. 2).
(Fig. 3). Direct confirmation that reaction with [5 - W ] methyl-Hr-folate yielded a ["C] methyl-B,, enzyme was obtained by extracting the chromophore from the protein with hot ethanol in the dark. It was subsequently identified as [W ] methyl-B,, by paper chromatography, paper electrophoresis, and its absorption spectrum before and after photolysis (68). The relative light stability of a [14C]methyl-B,, enzyme (Table IX and Fig. 3) was unexpected in view of the previously observed lability of a propyl-B,, enryme (2'7).Free methyl-B1, is rapidly degraded by light under aerobic conditions, the major cleavage product being formaldehyde (60).Therefore, a variety of solvent conditions were tested to determine if the cobalt-methyl group could be made light sensitive. Only acidification to below pH 2.5, which also precipitates the enzyme, resulted in complete photolysis (68). Photolysis at acidic pH values is not the result of removal of the entire cobalamin from the protein, followed by light cleavage of the freed [l4C]methyl-Bl2;it is only the subsequent exposure of acid-precipitated [ W]methyl-B,, enzyme to light which causes a 90-100% loss of "C from the precipitated protein (67,68, 6 1 ) . The light promoted loss of "C from the precipitated BI2protein was correlated with 60. H. P. C. H.ogenkamp, Biochemistry 5, 417 (1966). 61. R. T. Taylor and H. Weissbach, BBRC 27, 398 (1967).
4.
FOLATE METHYLTRANSFERABEB
141
Wovelength, nm
FIG.3. Effect of methylation with [5-"CImethyl-E-folate on the absorption spectrum of vitamin Blr transmethylase. (-) Vitamin &r enayme: initial enzyme, 22 mg/ml, in 0.05 M KHPO,, pH 7.4; (-- -) ["Clmethyl-Bu enzyme: same concentration of enzyme in K3HP04buffer after methylation with CB-l'Clmethyl-&folate (plus unlabeled AMe) followed by Sephadex G-25 filtration tmd lyophilization in the dark; eame tracing was obtained in the dark and after illumination for 20 min with a 100-W tungsten lamp at 10 cm and 0 " ; (--) ["Clmethyl-h ensyme + homocysteine: spectrum immediately after the addition of 0.2 pmole of homocysteine a t no,same tracing was obtained after an additional 3O-min aerobic incubation at 37". From Taylor and Weissbach (68).
the release of 14C-formaldehydeinto the supernatant fluid (68). An increased light stability of bound methyl-Blz vs. free methyl-B,, has also been observed when a [14C]methyl-B1,enzyme is prepared from apoenzyme + [14CImet,hyl-B,z(17). Shortly after a [14C]methyl-B,zenzyme was first reported to be relatively light stable (67),Stavrianopoulos and Jaenicke (96) offered a 1different explanation. These investigators also prepared a [**C methyl-B,, enzyme from radioactive folate substrate and a BIZ protein purified from E. coli. They attributed the retention of 6576% of its 14C after lighting (30 min, 10 cm, 160 W, 22') to a secondary reaction between the photolytically produced methyl radicals and amino acid side chains in the protein ($66).Such an explanation, however, cannot account for the findings made by Taylor and Weissbach (87, 68, 61).
142
ROBERT T. TAYLOR AND HERBERT WEISSBACH
FIQ.4. Structure of methylB,, (6,6-dimethylbenzirnidazolyl-cobamidemethyl).
First, photolysis did not alter the reactivity of their [W]methyl-B12 enzyme with homocysteine (67,68). Second, light did not alter its absorption spectrum (Fig. 3) which occurs when the [W ] methyl-cobalt bond is irreversibly broken. Third, lighting the [14C]methyl-B, enzyme did not decrease the yield of free ["C] methyl-BI2 which was subsequently extracted off the protein (68). Rudiger (62) has since confirmed independently that eneyme-bound ["C] methyl-B12 is indeed less light sensitive than free [14C]methyl-B,2. There are other examples in the literature of an increased carboncobalt bond stability to light. Sheep liver methylmalonyl mutase was purified as a light-stable 5'-deoxyadenosyl-B12-protein complex (63)and S- (9-adenyl) butyl-Blt forms a light-stable complex with ethanolamine deaminase (64). In a free state, both of these alkylcorrinoids are as photolabile as methyl-B,,. Pailes and Hogenkamp (66) found that the first-order rate constant at pH 7 for the photolysis of methylcobinamide was about 15% less than that for methyl-Blz. In the presence of imidazole, the cobalt in methylcobinamide appears t o become even more electrophilic because the photolysis rate constant decreases an additional twofold. The light stabilizing influence of imidazole is lost a t low pH 62. H. Riidiger, Eur. J . Bwchem. 21, 264 (1971). 63. J. J. B. Cannata, A. Focesi, Jr., R. Mammder, R. D. Warner, and S. Ochoa,
JBC 240,3249 (1966). 64. B. Babior, H. Kon, and H. Lecar, Biochemistry 8, 2662 (1960). @. W. H. Pailes and H. P. C. Hogenkamp, Bwchemktw 7, 4180 (1968).
4.
143
FOLATE METHYLTRANSFERASES
values. From such model experiments, it has been suggested that in the methylated B,, holoenzyme. the cobalt-5,6-dimethylbenzimidazolyllink is broken and a new bond between a histidine side chain and the cobalt is formed (66). In contrast, in the propylated BI2 holoenzyme, the histidine is not able to bind to the cobalt as a result of the added inductive effect of the extra ethyl group (66). While these are certainly plausible suggestions, it would seem that other possible contributions by the protein cannot be overlooked. Photolysis involves a homolytic split of the carbon-cobalt bond, and it requires oxygen (23,60). Consequently, apoenzyme binding may merely protect the methyl-cobalt bond from oxygen, and, in addition, stabilize the methyl radical, thereby favoring its recombination with cobalt (17, 68). Experimental evidence obtained with the use of methyl-B,, enzyme for any of these suggestions is presently lacking. Regardless of the light stability considerations, the substrate methylation experiments (Tables VII-IX and Figs. 2 and 3), a prior;, indicated that reaction (1) had been separated into two partial reactions (10) and (11). [5-W]Methyl-H4-folate
+ Blr enzyme FMNHi + 1,4-dithiothrsitol AMe
-
b
P4C]methyl-Bn enzyme
[W]Methyl-Blrenzyme
+ homocystaine aerobic
+ H,-folate
(10)
+
[14C]methylmethionhe Bn enzyme (11)
In reaction (lo), the enzyme acquires a cobalt-methyl group from the folate substrate, this step being dependent on AMe and a reducing system. In reaction (11) the methyl group is transferred to homocysteine, this step being closely analogous to reaction (8), although reaction (8) is catalyzed at a separate site (16) on the B, protein. However, it became apparent that the AMe-dependent accumulation of a ["C] methyl-B,z enzyme could not simply be described by reaction (10) when the process was examined stoichiometrically. It is seen in Table X that too Ilttle I-H,-folate is formed in the absence of homocysteine (complete system) to account for the amount of [14C]methyl-B1, protein that was produced in these same incubations. This finding and a better understanding of reaction (10) became clear when the function of AMe in reaction (1) was further elucidated. These points are discussed together in Section I1,E.
E. STUDIEs
ON THE
ROLEOF 8-ADENOSYL-L-METHIONINE
Early studies demonstrating the need for catalytic amounts of AMe in reaction (1) relied solely upon the use of relatively crude enzyme (14,
144
ROBERT T. TAYLOR AND HERBERT WEISSBACH
TABLE X FORMATION OF H,-FOJATE UPON METHYLATION WITH 1&"C1METHYL&-F0LATEa
Reaction mixture
Trichloroacetic acid precipitate (CPd
Completeb Complete! --FMNHs and platinum -BIS Enzyme -AMe +Homocysteine a
13,695 14,495 2,147 245 4,388
240
1.C Bound (mpmole) 0.18
0.20 0.03 0 0.06 0
a-Folate formed (mpmole) 0.038
0.038 0.016 0
0.02
0.43
From Taylor and Weissbach (4.9).
* Complete systems (0.2 ml) contained K&WOd, pH 7.4, 20 pmoles; [li-"C]methyl-
&-folate (73,000 cpm/mpmole), 1.0 mpmole; AMe, 1.0 mpmole; BIZenzyme, 0.92 mpmole; 1,44ithiothreitol, 5 pmoles; FMNHZ, 50 mpmoles; and platinum, 0.1 mg. After 5-min incubations in the dark at 37"under HI gas, 0.3 ml of cold water were added to each mixture. Samples of 0.2 ml were t8henremoved for trichloroacetic acid precipitation and Millipore filtration. To t,he remaining 0.3 ml of the diluted mixtiires was added 0.1 ml of potassium ascorbate, pH 6.0,100 mg/ml. These ascorbate-protected samples were stored at - 15" in the dark and then assayed microbiologically for H.-folate.
S4,Sb). They provided no information as to how AMe functions catalytically except that it was shown (9, 34) that AH could not be substituted for AMe. Prompted by the observation that both AMe and methyl iodide could prevent chemical propylation (b7), methyl iodide was tested and found to satisfy partially the requirement for AMe in reaction (1) (66). Methyl iodide was also shown to function catalytically in the system (67). Since radioactive folate substrate was used in these experiments, any direct nonenzymic reaction between the unlabeled methyl iodide and homocysteine did not affect the estimation of [5-I4C]methyl-H4-folatehomocysteine transmethylation. These observations with methyl iodide indicated that AMe could only be serving as a methyl group donor. Since the slow catalysis (Table V) of AMe-homocysteine transmethylation [reaction (6)] was inhibited by chemical propylation (27), a possible site for methylation by AMe (or methyl iodide) was t.he BIZcobalt. It is known that methyl iodide (69, 68) and AMe (69) will react rapidly with free B,,, to form methyl-Blz. 66. R. T. Taylor 67. R. T. Taylor 68. E. L. Smith, lSa, 1175 (1962). 69. W.Friedrich
and H. Weissbach, JBC 241, 3641 (1966). and H. Weissbach, JBC W , 1517 (1967). L. Mervyn, A. W. Johnson, and N. Shaw, Nature (London) and E. Kiinigk, Biochem. Z.336, 444 (1962).
4.
145
FOLATE METHYLTRANSFERASES
Direct evidence for AMe methylation of the cobalt was obtained by using [ "C J methyl-AMe (Table XI), A reducing system was required and the properties of the resulting ["C]methyl-B,, enzyme (light stability and homocysteine reactivity) were indistinguishable from those of ensyme that had been methylated with [5-14C]methyl-H,-folate (plus unlabeled AMe) (49,61). Consequently, it was anticipated that both unlabeled 5-methyl-H4-folate and homocysteine should markedly decrease the accumulation of a ["C] methyl-B,, enzyme from ["C] methyl-AMe. This, in fact, did happen as seen in Table XI. Homocysteine very likely functioned as a methyl group acceptor and demethylated most of the [14C]methyl-B,, enzyme to form [14C]methylmethionine, but the effect of unlabeled 5-methyl-H4-folate on the methylation of the enzyme by ["C] methyl-AMe was not clear and required further study. Short-time enzyme labeling experiments (Fig. 5) revealed that methylation by [14C]methyl-AMe occurred within the first 30 sec of incubation. I n the absence of unlabeled 5-methyl-H4-folate the ["Clmethyl from ["C] methyl-AMe remained on the enzyme. But in the presence of unlabeled folate substrate, it was removed from 30 sec to 6 min after the start of the incubation. In a correlative experiment (Fig. 5), the AMe-dependent formation of [' C ] methyl-B,, enzyme with [5-14C]methyl-H4-folate showed a lag during the first 30 sec and required 5-6 min to reach completion. The data in Fig. 5 indicated that in the absence of homocysteine a [14CJ methyl-Bln enzyme accumulated from [5-"C] methyl-It-folate by an exchange with a cobaltmethyl group derived initially from unlabeled REQUIREMENTS
TABLE XI
FOR THE
FORMATION OR
TRICHLOROACETIC
ACID-PRECIPITABLE l4C
WITH [ 1 4 C ] M ~ ~ ~ b A M e 4
Reaction mixture Complete (37")b -BIZ Enzyme -FMNH* and platinum +bMethyl-H4-folate +Homocysteine
Trichloroecetic acid precipitate (cpm) 12,800 1,480
2 ,074 2,600 2,470
14C Bound (mlunole) 0.61 0 0.03 0.06 0.05
From Taylor and Weissbach (49). Complete systems (0.2 ml) contained KzHPO,, pH 7.4, 20 e o l e s ; [14C]methyl-AMe (18,500 cpm/mpmole), 10 mpmolea; 1,4dithiothreitol, 5 jmoles; BH enzyme, 0.54 mpmole; FMNH,, 50 mpmolea; and platinum, 0.1 mg. Where indicated dL%methyl-&folate, 30 mpmoles, or homocysteine 0.2 pmole, waa added to the complete system. Incubations were for 15 min in the dark under HI gas. a
b
146
ROBERT T. TAYLOR AND HERBERT WEISSBACH
Minutes
FIG.5. Time dependence of [l’Clmethyl-Blr enzyme formation with [“ClmethylAMe and [5-14Clmethyl-&-folate. All reaction mixtures (0.2) ml contained 0.46 mpmole of Bu enzyme; 10 mpmoles of AMe either unlabeled or [“Clmethyl (18,000 cpm/mpmole) ; and, where indicated, 10 mpmoles of dZ-6-methyl-EL-folate either unlabeled or [“Clmethyl (73,000cpmlmpmole) . Incubations were under HZ gas at 37” for the times indicated. They were initiated within 3 min after the injection a t 0” of reduced flavin. From Taylor and Weiasbach (49).
AMe. Thus, the dependency on AMe for methyl transfer from [5-14C]methyl-Ha-folate to the enzyme (Table VII) and the formation of insufficient Ha-folate (Table X) would be explained by this exchange reaction. The mechanism for this exchange was realized when analysis showed (49) that the dl- [5-14C]methyl-H,-folate prepared by the method of Keresztesy and Donaldson (IS) contained traces of dl-Hafolate. Enzyme that had been methylated with [14C]methyl-AMe was shown to react with HI-folate to produce the active isomer of [5-”C]methyl-Ha-folate (@). Based on these findings the mechanism for the results in Fig. 5 is depicted in Fig. 6. H4-Folate accepts a methyl group (derived initially from AMe) from methyl-B,, on the enzyme to form 5-methyl-HI-folate and a postulated Col+ enByme species (BIZsenzyme)
4.
147
FOLATE METHYLTRANSFERASES
H,-Folate TZ.9-3-6 (fe)
N6-Methyl-H,-folate Enz
FIQ.0. Role of a-folate in methyl group exchange between 6-methyl-&-folate and the methyl-Btt enzyme. From Taylor and Weiesbach (81, 49).
The latter then reacts reversibly with exogenously added [5-I4C]methylH,-folate to reform a [14C]methyl-B,, enzyme with the methyl group now having been derived from the radioactive folate substrate ( 3 1 ) .Figure 6 predicts (a) that small amounts of H4-folate should accelerate the AMe-dependent rate of [14C]methyl-B,, enzyme accumulation from an excess of [5-14C]methyl-H,-folate, (b) the B12enzyme should catalyze AMe-H,-folate transmethylation [reaction (7) 1, and (c) the B,, enzyme should catalyze an AMe-dependent exchange reaction between pteridine ring labeled 5-methyl-H4-folate and unlabeled H4-folate to give pteridine ring labeled H4-folate. All of these predictions have been demonstrated experimentally (31). The above results on exchange transmethylation (31, 49) suggested a mechanism for the participation of AMe in reaction (1). Inactive B12 protein is first primed or activated by B,,-cobalt methylation. This priming can be facilitated by AMe or methyl iodide. Once the enzyme has been converted to one of its functional forms by AMe methylation, the enzyme should catalyze reaction (1). The key to testing for AMe activation depended on devising an experimental system that ‘would permit one to detect even a few cyclic turnovers by the enzyme. Table XI1 summarizes the results of such an experiment. Millimicromole amounts of B,, enzyme were first incubated with AMe (or methyl iodide) in the presence of an FMNH, reducing system. The reaction mixtures were then shaken aerobically to oxidize the flavin and reincubated with the two substrates, 5-methyl-H,-folate and homocysteine. The activated enzyme (methylated in the first incubation) catalyzed the aerobic synthesis of about 10 equivalents of [14C]methylmethionine (Table XII-) . The catalysis of reaction (1) was complete in 30 sec because the enzyme was approximately two-thirds inactivated within 5 sec after adding the substrates (31).Requirements for both AMe and FMNH, during the first incubation were also observed when the activated enzyme was separated from the smaller components in the preliminary reaction mixture, prior to being challenged with the two substrates (31).The important point
148
ROBERT T. TAYLOR AND HERBERT WEISSBACH
TABLE XI1 REQUIREMENTS FOR AEROBIC METHYL[W]GROUPTRANSFER FROM [5-"C]METHYLH~-FOLATETO HOMOCYSTEINE~ First incubation reaction mixture Compleh? Complete (FMN) -FMNH* and platinum -AMe -AMe -AMe CHJe -BIS Enzyme Complete
+
15 min, 37" Gas phase
HI Air Hz HI HZ Ha Hz HI
[W]Met,hylmethionineformed after a second aerobic incubation, 5 min, 37"~(mpmole) 3.4-3.8 0.07 0.01 0.10 0.13 (+AMe)d 3.40 0
0.09 (- homocysteine)f
From Taylor and Weissbach (31). *Complete systems (0.2 ml) for the first incubation contained Blr enzyme, 0.39 mpmole; K*HP04, pH 7.4, 20 pmoles; AMe, 10 mpmoles; FMNHz, 50 mpmoles, and platinum 0.1 mg. c A t the end of the first incubation the reaction mixtures were chilled to 0" in the dark and shaken for several minutes to oxidize the FMNHz. They were then equilibrated a t 37" for 5 min and tested for their transmet.hylaseactivity by the addition of 0.06 ml of a pH 7.4 substrate solution containing 0.5 mM dl-[5-W!]methyl-H4-folate (14,400 cpm/ rnpmole), 2.5 mM homocysteine, 0.05 M l+dithiothreitol, and 0.05 M KzHPO~, pH 7.4. ["CIMethylmethionine synthesis was determined a t the end of a 5-min incubation (16). d AMe (10 mpmolea) was added at the end of the first incubation after the FMNHz had been oxidized. "Methyl iodide (0.15 pmole in 2 pl of ethanol) was subst,ituted for AMe and the complete system (0.2 ml) also contained 1,4ditmhiothreitol(5 pmoles) during the first incubation. Homocysteine was omitted from the second incubation system.
in these experiments was that limited ["C] methylmethionine synthesis occurred only when the first incubation conditions were such as to permit the preliminary formation of a methyl-B,, enzyme (49, 68, 6 1 ) . Thus, both AMe (or methyl iodide) and reduced flavin were always essential (31) to preform a catalytically active B1, protein. It has been verified with the use of apoenzyme that only cobalt methylation is involved in AMe activation. Purified B,, protein was resolved with urea into apoenzyme (17)as discussed earlier (Section I1,B). The apoprotein was incubated with [l*C]methyl-B,, and then freed of unbound ["C] methyl-B,, with a charcoal column. Since this [ "C] methyl group reacts quantitatively with homocysteine (17),one has an accurate estimate of the amount of reconstituted [ "C] methyl-B,? enzyme added to a system. The importance of testing its catalytic activity resided in the fact that it was a methylated B,, enzyme that had been prepared by cir-
4.
149
FOLATE METHYLTRANSFERAGES
cumventing a preliminary exposure (Table XII) to AMe and FMNH,. Hence, aerobic turnover (Table XII) resulting from activation by any reduced flavin or by nonspecifically bound AMe (49) could be eliminated. Table XI11 demonstrates that reconstituted ["C] methyl-B,, enzyme (0.25 mpmole) synthesized aerobically 18 equivalents (4.5 mpmoles) of [14C]methylmethionine in the absence of both AMe and FMNH2. The activation of the enzyme was not lost if the folate substrate was added prior to homocysteine, but the catalytic activity was lost by demethylation with homocysteine for 15 sec prior to addition of the [5-14C]methylH,-folate. A similar observation was made using enzyme that had been activated by AMe methylation ( 3 1 ) , i.e., an activated enzyme was a methylated enzyme, and any procedure that removed the methyl group resulted in complete loss of activation. These findings were indicative of an extremely oxygen-sensitive form of the cobalamin (such as B12*, Fig. 6) being formed in the catalytic cycle. Consistent with this view, it was observed that the total catalytic turnover of both AMe premethylated (31) and reconstituted (32) methyl-B,, enzyme was increased as much as 10-30-fold by carrying out reaction (1) under anaerobic conditions (e.g., H, gas phase, Fig. 7). Under H, gas, the catalytic life of reconstituted [14C]methyl-B,, enzyme lasted sufficiently longer to increase the total number of enzyme turnovers to 530-fold. A turnover rate (after 5 sec) of 910 molecules of [14C]methylmethionine formed per minute per molecule of added methyl-B,, enzyme was calculated from the data in Fig. 7. Like an AMe-premethylated enzyme (N), the reconstituted enzyme in Fig. 7 could not maintain its initial rate of synthesis unless both AMe and FMNH? were included in the system (39). Irrespective, the initial turnover rate in Fig. 7 is comparable to the steady-state turnover TABLE XI11 EFFECT OF ORDEROF SUBSTRATE ADDITION ON T F ~ EAEROBIC CATALYSIS BY RECONSTITUTED (WIMETHYGBI~ ENZYME^.^ Order of addition
+
[5J4C]Methyl-H4-fo1ate homocysteine (5 min) (5-W]Methyl-H4-folate(15 sec), then homocysteine (5 min) Homocysteine (15 sec), then [5-14C]methyl-H4-folate(5 min)
A
P'C1Methylme thionine formed (mmole) 4.5 4.0 0.4
From Taylor and Henna (96). [W]Methyl-BI1enzyme, 0.92 mg (0.25 mamole) in 0.2 ml of 0.1 M KnHPO4, pH 7.4, was equilibrated at 37" for 5 min. A 1.0 mM solution of d~-[-[5-14C]methyl-H~-folate (29,000 cpmlmpmole), a 5.0 mM solution of homocysteine, and a 1: 1 mixture of the mbstrate solutions were also equilibrated separately at 37"for 5 min. Then 0.06 ml of the 1 :1 mixture or 0.03 ml of each substrate was added and incubated at 37" a~ shown. a
b
150
ROBERT T. TAYLOR AND HEEBERT WEISSBACH
7.0I
I
I
I
I
6.0
f
u)
5.0
g' 3.0 t 0 ..-
f
2.0
f *"
1.0
0
-I
Reconstituted mathyl-"C-B,, Holoenzym 12.5 pmoles, minus homocvsteine (m)
I
I
I
I
I
0
1.0
20
3.0
40
1 5.0
Minutes, 37'
FIG.7. [5-14C1Methyl-&-folate-homocysteine transmethylation under HZ gas by C'Clmethyl-&a holoenzyme. Final reaction mixtures (0.26 ml) contained K H P O , pH 7.4, 20 pmoles; dZ-[5-"CImethyl-H,-folate (29,000 cpm/ mpmole-, 30 mpmoles; homocysteine, MI mpmoles; and enzyme, either 46 p g of reconstituted C"C1methyl-Bu enzyme (12.5 pmoles of bound C1'CImethyl-Bl2) or 50 p g of the original Bu holoenzyme (50 pmoles of bound Baa). Enzyme in 0.1M KzHP04buffer (0.2 ml) and a pH 7.4 substrate solution containing 0.5 mM dl-[5-"Clmethyl-a-folate plus 2.5 mil4 homocysteine were pregassed separately with H, for 5 min at 0' and with constant agitation. Then the buffered enzyme and substrate mixture were equilibrated separately a t 37' for 5 min. Reactions were initiated by the injection of 0.06 ml of substrate mixture and were terminated a t the times indicated with 0.8 ml of ice cold water. From Taylor and Hanna (3.8).
a reconstituted
rate of 860 which was obtained using the reconstitued enzyme and conditions of Fig. 7 except with the supplementation of AMe FMNH, (32). It was, therefore, concluded that a methyl-B,, protein is a fully active enzyme species which, a t least transiently, requires only the two substrates to catalyze reaction (1); AMe and reduced flavin are only involved in the preliminary formation of methylated enzyme (32). Rudiger and Jaenicke (70,71) have since reported the partial purification of a methyl-B,, holoenzyme directly from extracts of E . coli. It is active in the absence of exogenously supplied AMe but only in a reducing
+
70. H. Rudiger and L. Jaenicke, Eur. J . Biochem. 10, 557 (1969). 71. H. Riidiger and L. Jaenicke, Eur. J. Biochem. 16, 92 (1970).
4.
151
FOLATE METHYLTRANSFEBASEB
+
+
system formed by adding lJ4-dithiothreitol aquo-B,, FMN. In an enzymic FADH, generating system, AMe is still essential (70). Also, in the AMe-independent reducing system, their methyl-B,, enzyme regularly ahows an unexplained lag in the catalysis of reaction (1) during the first 15 min of incubation (70).It should be noted that the absorption spectra of the latter enzyme preparations (70, 71) differ significantly from that of free methyl-B, (69) or that of the methyl-B,, enayme prepared in vitro by other investigators (17, 68). Thus, before assessing their data, one should wait until the cobalamin is removed from the protein and adequately characterieed as methyl-B,,. It is possible that by avoiding all precipitation-type steps ($9, 70, 71), their new enzyme may show AMe-independent activity as a result of the presence of ionically bound, endogenous AMe (49). The extended lag in AMe-independent catalysis (70) might reflect the slow release of this endogenous AMe in a 1,4-dithiothreitol-containing reducing system.
F. MECHANISM OF N 6 TRANSMETHYLATION
-
M
~
~
A schematic mechanism for reaction (1) that is compatible with the data presented (Sections II,C,D, and E) is shown in Fig. 8. It is not intended as a detailed kinetic scheme to account for all of the possible substrate and product complexes with the enayme that may exist. Hence, even the now-known complex (4.8) between B12enByme and the intact l-5-methyl-H4-folate substrate has been deleted. The usefulness of Fig. 8 for discussion purposes is that it shows the relationship between the three forms of bound BI2 that have been identified with the enzyme. Figure 8 also accounts for the equal sensitivity of reactions (1) and (5)-(7) to chemical propylation and the catalysis of exchange transmethylation. The salient feature when this scheme was first proposed (31) was that it
Ns-Methyl-4-folate
Methlonine
~
~
152
ROBERT T. TAYLOR AND HERBERT WEISSBACH
assigned a cobalt methylation role to AMe. Yet, i t satisfied the fact that unlabeled 5-methyl-H,-folatc does not inhibit (9, 16) the slow catalysis of reaction (6). In Fig. 8, the notation S-
represents collectively any or all of the nonmethylated but inactive forms of the B,, chromophore that may exist, including the salmon-colored derivative that is customarily associated with purified preparations (16, 29).Although the precise structure of the salmon-colored cobalamin is uncertain, the above symbol does denote the fact that a sulfur-containing B,, compound has been recovered (19,22) from most of these preparations. Irrespective of its exact structure, the salmon-colored protein isolated by typical purification methods (16,29) clearly does not contain methyl-B,, (19,22). The first step in the sequence is methylation of the inactive enzyme by AMe (or CHJ) in the presence of a reducing system. No attempt is made (Fig. 8) to indicate the number of steps involved in the conversion of the initial B12 protein to a methyl-B,, enzyme. It may be worthwhile, however, to comment on the role of the reducing system in this reaction. Reduced flavin has the redox potential to reduce aquo-B,, or cyano-B,, by at least one electron to BIZ, (79).It also undoubtedly reacts with traces of oxygen to improve the anaerobiosis of a system. Reduction of the bound B,, to the Co2+ (B,,,.) stage followed by disproportionation (73) might produce traces of B,?, enzyme aquo-B,, enzyme. Although the equilibrium of disproportionation greatly favors BIzr (73), it has been suggested (63)that AMe (or CHJ) with its high methyl group transfer potential (74) can trap such traces of BIznenzyme. Disproportionation is attractive in that the bound B,, could be methylated initially without a direct two-electron reduction of the cobalt. This would avoid overcoming the E’,, of -1.07V (7.2) required to reduce B,,, to BltR.Howcvcr, disproportionation (73) requires that the chromophores on two separate molecules of BIZ,.protein must react. Conclusive spectral evidence showing that FMNHz or FADH, per sc (in the absence of thiols) will reduce an inactive B,, protein to a BIZ, protein has not yet been obtained. It is clear, though, that whatever the form of this reduced B,, protein prepared from the inactive enzyme, it has a preference for AMe, not 5-methyl-H,-folate.
+
72. H. P. C. Hogenkamp and S. Holmes, Biochemistry 9, 1886 (1970). 73. R. Yamada, S. Shimizu, and S. Fukui, Biochemistry 7, 1713 (1968). 74. S. H. Mudd, W. A. Klee, and P. D. Roea, Biochemistry 5, 1653 “66).
4.
FOLATE METHYLTRANSFEBASES
153
I n the next step in Fig. 8, the methyl-B,, enzyme formed by AMe methylation of the inactive enzyme transfers its methyl group to homocysteine to form methionine plus a Cox+ enzyme. The latter species is then methylated by the folate substrate, and all subsequent methyl groups appearing in methionine come from the folate substrate. Thus, the methyl group from AMe is used only to activate the enzyme. Although AMe and 5-methyl-H4-folate can both methylate the B,, cobalt, these reactions take place sequentially on different species of the BIZ protein. In 5-methyl-H4-folate transmethylation, two forms of the bound cobalamin (methyl-B,, and B,,,) function cyclically as prosthetic groups analogous to pyridoxamine phosphate and pyridoxal phosphate in enzymic transamination. While the existence of a methyl-B,, enzyme is well-documented (Sections II,D and E), the evidence for a Col+ enzyme (B,,, enzyme) pictured in Fig. 8 has been largely circumstantial or indirect. The initial evidence was based on the extreme oxygen lability of the Blz enzyme during transmethylation (31, 38) in comparison to the known nucleophilicity and lability of free B,,, (21, 93, 54, 75).Recently, however, spectral evidence for a Col+ enzyme (BIZ, enzyme) was obtained in an anaerobic system containing methyl-"C-AMe, homocysteine, He-folate, l,Cdithiothreitol, and the salmon-colored B12 protein (76).During transmethylation from methyl-14C-AMe to homocysteine [reaction (6) 1, the B1,,.-like spectrum of the salmon-colored protein shifted to give new absorption maxima at 385 and 460 nm. These maxima are highly indicative of B,,, formation (77).They were transient, however, even under H, gas, since the B,,, spectrum reverted back to a B,,, spectrum as soon as the methyl-"C-AMe had been metabolized (76).Therefore, as seen in Fig. 8, the Col+ enzyme is pictured as continually oxidizing, even in a highly anaerobic system, to an inactive BISprotein that must again be primed (methylated) by AMe. One can regard the slow methylation of homocysteine by AMe [reaction (6)] via a methyl-BIZ enzyme (37, 31, 4.9) as a manifestation of its cofactor role in reaction (1). In support of these views is the observation that the apparent K,,, for AMe in reaction ( 1 ) increased from 5.2 X WRM to 5.0 X lo-' M (32)upon 1,4dthiothreitol reducing system to a switching from an FMNH, 2.5-fold-less active reducing system (15).This is what one would predict from Fig. 8 since the apparent K,,, for AMe should largely reflect the rate of Colt enzyme inactivation in a particular system.
+
75. G. N. Schrauaer, E. Deutsch, and R. J. Windgaesen, J . Amer. Chem. SOC.90,
2441 (1968). 76. R. T. Taylor and M. L. Hanna, BBRC 38, 7Ei8 (1970).
77. R. Bonnet, Chem. Rev. 63, 573 (1963).
154
ROBERT T. TAYLOR AND HERBERT WEISSBACH
Unfortunately, fnr both the B,?-dependent [reaction (1) ] and the B,?-independent [reaction (2) ] synthesis of methionine, no concrete information is available as to how the N6-methyl group is activated for enzymic transfer. 5-Methyl-H4-folate is not a high energy onium compound like AMe (74).Instead, its N6-methyl group is chemically quite unreactive a t pH 7, even with free BIZs (68, 78). Recently, it was reported that a Rlight amount of nonenzymic reaction takes place in acidic media between 5-methyl-H4-folatc and B,?, to yield a trace of methylBiz (6%. 111. Non-B,, Methyltransfemse from Gcher;ch;a col;
A. ASSAYAND PURIFICATION Non-B, transmethy lase activity can be routinely assayed by measuring the formation of [14C]methylmethionine using [5-14C]methyl-H4folate (GluJ as a substrate. The radioactive methionine can be separated from the folate substrate with an anion exchange resin (14). A typical reaction mixture would contain in a total volume of 50 pl: dE- [5-"C] methyl-H,-folate (Glu,) , 3 mpmoles, 13,300 cpm/mpmole ; L-homocysteine, 50 mpmoles; Na,HP04 buffer, pH 7.8, 500 mpmoles; magnesium acetate, 5 mpmoles ; lP-dithiothreitol, 500 mpmoles ; and enzyme. After incubation for 15 min a t 37", the reaction is stopped by TABLE XIV
PURIFICATION OF NoN-BI~ TRANSMETHYLASE FROM E. w2i KIP
htion
Volume (ml)
Total activity (k-unit4
1. Extract 2. Ammonium sulfate (0435%) 3. Protarninesulfate 4. Ammonium sulfate (50435%) 5. Fimt DEAE-Sephadex 6. Hydroxylapatite 7. Second DEAESephadex 8. SephsdexG-100
319 154 306 31 102 61 85 47
1220 1080 1080 664 625 376 362 320
Specific Protein activity (mdml) (units/me) 30.4 50.0 15.4
48.4
4.28 3.58 2.03 2.7
126 140 232 463 1430 1720 2100 2520
Yield
(%I
100 89 89 54.5 51 31 29.7 26.2
From Whitfield et at. (79). 78. G. N. Schrauzer and R. J. Windgaesen, J . Amer. Chem. Soe. 89, 3607 (1967). 79. C. D. Whitfield, E. J. Steers, Jr., and H. Weissbach, JBC 245, 390 (1970).
4.
FOLATE METHYLTRANSFERASES
155
the addition of 0.9 ml of cold water. The [14C]methylmethionineis then separated from the unreacted [ 5-"C] methyl-H4-folate (Glu,) by chromatography on a Dowex 1 (Cl-) column (0.5 by 3 cm) and assayed for radioactivity (79). The enzyme has been purified from E. coli AB1909, a methionine auxotroph defective in N6Jn-methylenetetrahydrofolate reductase by Whitfield et al. (79).Undcr derepressed conditions the enzyme represents about 5% of the soluble protein of the cell. A summary of its purification is seen in Table XIV. One unit of non-Blz transmethylase activity is defined as the amount of enzyme catalyzing the formation of 1 mpmole of ["C] methylmethionine per 15 min at 37" under standard assay conditions. A homogeneous material has been obtained, and the protein has been partially characterized. B. PHYSICAL PROPERTIES The sedimentation coefficient, S q W , of the enzyme was determined from velocity ultracentrifugation a t a protein concentration of 10 mg/ml. It was calculated to be 4.7 S (79) by the method of Schachman (80). The molecular weight of the enzyme was determined to be 84,OOO by equilibrium sedimentation according to the meniscus depletion method of Yphantis (81). Similar experiments employing a carboxymethylated transmethylase in the denaturing agent, 5 M guanidineOHC1, revealed a lower molecular weight of 50,800. The lower molecular weight of the denatured, carboxymethylated enzyme suggests that the native transmethylase is composed of two subunits. The percentage of nitrogen in the transmethylase was found to be 16.7, and the gram per liter extinction coefficient, E 3 , of the enzyme at 280 nm in 1 X 10-2M, N&HP04, pH 7.8, was 1.62 liters g-' cm-l. In 0.1 M NaOH, the E was 1.6 liters g-l cm-l a t 280 nm, and the Eg was 1.18 liters g' cm-I at 294 nm. The latter values were used to calculate the tryptophan and tyrosine content of the protein. Analysis of amino acid composition revealed that the most frequently occurring amino acids were aspartic acid, glutamic acid, alanine, and leucine. Cysteine and methionine were present in the lowest amounts. Using an enzyme molecular weight of 84,000 and a protein concentration based on dry weight and nitrogen analysis, the tyrosine residues per molecule were 23.2 from spectrophotometric determination, as compared to 18.9 from amino acid analysis (79). The number of tryptophan residues per 80. H. K. Schachman, "Methods in Enzymology," Vol. 4, p. 32, 1957. 81. D. A. Yphantis, Biochemistty 3, 297 (1964).
156
ROBERT T. TAYLOR AND HERBERT WEISSBACH
molecule by spectrophorometric determination and spectrophotofluorometric analysis was 18.4 to 20.4.
PROPERTIES AND BINDING OF FOLATE SUBSTRATE C. CATALYTIC Characteristics of the overall catalytic reaction [reaction (2) ] were examined using a homogeneous preparation of the purified enzyme (79). The formation of [ "C] methylmcthionine from [5-14C]methyl-H,-folate (Glu3) was dependent on the presence of homocysteine, enzyme, and phosphate and was stimulated by Mg2+and 1,4-dithiothreitol as seen in Table XV. The Na2HP0, requirement was specific for the phosphate ion since KzHP04could replace Na2HP0,, but C1- and sulfate ion could not. Magnesium ions could be replaced by manganese ions and less effectively by calcium ions. The enzyme was nonspecifically inhibited by a high ionic strength medium. TABLE XV
RBQUIREMENTB OF NoN-B~,TRANSMETHYLASE FOR METHIONINB SYNTHESISO Methionine formed Reaction mixture Complet& Homocysteine -Enzyme - N d P O 4 buffer -N%HFQ4 buffer -N&HPO4 buffer -N&HP04 buffer
(cpm)
(pmoles)
2715 0 0 328 2685 280 220 1331 0
202 0 0 24.3 199 20.8 16.3 98.6 0
-
-Me+
+ KsHP04 buffer + NaCl + tris-Cl buffer
- [5-W]Methyl-H4-folate (Glur) + [5-14ClMethyl-H,folate (Glul) or [51'C]Methyl-&-folate (a-Glur) -Homocysteine + cysteine, 2-mercaptoethanol, or
0
0
1,4-dithiothreitol l,&Dithiothreitol
+
3017 ~
~
~~
224 ~
From Whitfield d al. (79). complete system contained dZ-[5-14C]methyl-H4-folate (Glu:), 3 mpmoles, 13,500 cpm/mpmole; chomocysteine, 50 mpmoles; Na3HPO4 buffer, pH 8.2, 500 mpmoles; magnesium acetate, 5 mpmoles; and purified enzyme, 0.12 pg, in a total volume of 50 pl. The following were added where indicated: 23 mpmoles of dE[5-l'C]met,hyl-H4folate (Glu]), 54,000 cpm/mpmole; 30 mpmoles of db[5-W]methyl-Hlate (a-Glui), 2340 cpm/mpmole; and 500 mpmoles of NaCl, KzHP04 buffer, pH 8.2; t,ris-Cl buffer, pH 8.2; l,rl-dithiothreitol, 2-mercaptoethanol, and ccysleine. The pH (8.1) of the reaction mixture did not change when N&HP04 buffer was omitted because of the buffering capacity of the other reaction components. 4
5 The
4.
FOLATE METHYLTRANSFERASES
157
The substrate specificity of the enzyme was investigatcd by the use of derivatives or analogs of the two substrates (79).[5-14C]Methyl-H4folate (Glu,) and [5-14C]methyl-Ha-folate (a-Glu,) could not replace the triglutamate folate derivative as a methyl donor. I n addition, cysteine, 2-mercaptoethanol, and 1,4-dithiothreitol could not replace homocysteine as a methyl acccptor. At pH 7.8 the K,,, for Z-[5-14C] (79). Under saturating conditions methyl-H,-folate (Glu,) was 2.35 and based on one catalytic site per enzyme molecule (88),the turnover number of purified transmethylase [reaction (2)] is only 14 mpmoles of methionine formed per minute per millimicromole of enzyme. Thus, compared to the E . coli B,, enzyme [reaction (1) turnover number about 800, Table V], the non-B12 transmethylase is a considerably less efficient catalyst. The pH optimum of the enzyme was determined in phosphate buffer only, since phosphate is required for the reaction. Other buffers such as tris, N-tris, (hydroxymethyl) methylglycine, barbital, and imidazole plus optimal levels of phosphate resulted in an inhibition of the reaction as a result of the high ionic strength created. The enzyme was active between pH 6.0 and 8.5, with optimal activity a t pH 7.5-7.8. Although no evidence for a substrate-methylated, non-Bt2 enzyme has been obtained, [5-14C]methyl-H4-folate (Glus) does form a specific complex with this enzyme in the absence of homocysteine (88).As seen in Fig. 9, incubation of the enzyme and dl- [5-14C]methyl-H4-folate (Glu,) a t 0" followed by chromatography on Sephadex G-50 yielded a peak of radioactivity which eluted with the protein. No radioactivity was seen in this region when the 14C substrate was chromatographed alone in a similar manner. Figure 9 also shows the highest level of enzyme eluted in fraction 12, whereas the highest concentration of complexed radioactivity occurred in fraction 13. The trailing of radioactivity in fractions 16-19 behind the cnzyme-substrate complex peak, as well as the slight displacement of the radioactive peak from the enzyme peak, suggested that the complex dissociated during chromatography. Studies on the nature of the bound radioactivity using enzymic and paper chromatographic procedures (82) showed that the active isomer of [5-14C]methyl-H,-folate (Glu,) was present intact in the complex. A Sephadex G-50 column in which the enzyme and substrate were constantly in equilibrium was used to study the stoichiometry of the binding of substrate to enzyme as well as t o ascertain the dissociation equilibrium constant of the cnzyme-substrate complex (88). The maximum amount of substrate bound corresponded to 0.76 mpmole of [5-14C]82.
C. D. Whitfield and H. Weissbach, JBC 245, 402 (1970).
158
ROBERT T. TAYLOR AND HERBERT WEISSBAClI
Fraction number
WQ.9. 1solat.ion of preformed C"C1enzyme complex by Sephadex
G-50 chromatography. After incubation of 0.6 mg of non-Bo transmethylase With 6 mpmoles of dl-C6-"C1methyl-H4-folate (Glud, 76,000 cpm, a t O", the mixture was applied to a Sephadex G-60 column: (0)radioactivity in the absence of enzyme, ( 0 ) d o a c t i v i t y after incubation with eneyme, and (A) absarbance of the enzyme a t 280 nm. From Whitfield and Weissbach (89).
methyl-H,-folate (Glus)/mpmole of transmethylase, using a molecular weight of 84,OOO (79)for the enzyme.
D. REPRESSION OF ENZYME SYNTHESIS In E . coli it is wcll established that methionine represses all of thc enzymes involved in its biosynthesis including the non-B1., transmethylase (79,83-85). In addition, it has been shown that cyano-BI2 also represses reductase the non-B,, transmethylase (86) and N5Jo-methylene-H4-folate (84). The available data (86) indicate that the repression of the non-Bln transmethylase observed with the vitamin does not result from excess synthesis of methionine as might be expected since cyano-B12 in vivo is metabolized to cobalamins which convert the apoenzyme form of the BE 83. R. J. Rowbury and D. D. Woods, J . Gen. Microbiol. 24, 129 (1961). 84. H. M. Katzen and J. M. Buchanan, JBC W , 826 (1965). 85. R. J. Rowbury and D. D. Woods, J . Gen. Microbiol. 35, 145 (1964). 86. L. Milner, C.Whitfield, and H. Weissbach, ABB 133, 413 (1969).
4.
159
FOLATE METHYLTRANSFERASES
TABLE XVI EFFECTOF VARIOUS COBAMIDE COMPOUNDS AND L-METHIONINE ON THE LEVELSOF THE Non-Blr TRANSMETHYLASE'
0
Addition to the growth medium
Specific activity
Repression
None L-Methionine 1 0 1 M L-Methionine lo-* M GMethionine 5 x 10-6 M L-Methionine 10-6 M Cymo-Bia 10-8 M Cymo-Bia 10-'0 M Factor B 10-7 M Factor B 10-8 M Factor B 10-0 M Cyano-Blranilide 10-7 M Cyano-B1panilide 10-8 M Cyano-Blranilide 10-8 M
13.5 2.2 2.6 8.7 12.1 0.1 12.8 10.1 10.2 13.4 12.2 12.3 13.6
0 83 80
(%I
35 10
99 5 5 23 0 11 10 0
From Milner et al. (86).
transmethylase to the active holoenzyme (87). The effect of L-methionine and several cobamides on the repression of the non-BI2 transmethylase in E . coli K,, is seen in Table XVI. Although both L-methionine and cyano-B12 repress synthesis of the non-BI2 transmethylase, the vitamin is more effective, and the repression by the various cobamides tested appears to correlate with the ability of the cobamide to form the Bl2 transmethylase holoenzyme. More recent studies using methionine auxotrophs of E . coli have provided further evidence that the mechanism of non-B,, enzyme repression by L-methionine and .cyano-B1, are different (88). In a methionine regulatory mutant (Met J-), L-methionine is not able to repress the enzymes involved in the biosynthetic pathway although cyano-Blr 7s still an effective repressor. In contrast, studies with an E . coli auxotroph lacking the BIZtransmethylase (Met H-) have shown that the non-Blt transmethylase is not repressed by cyano-BI2, but it can be repressed by L-methionine (88). The above results indicate that the vitamin and other cobamides capable of forming the B,, transmethylase holoenzyme can specifically repress the synthesis of the non-B1, transmethylase. The B,, enzyme itself appears to be part of the repressor system, but the regulatory gene required for repression of all the biosynthetic enzymes 87. H. Weissbach, B. G. Redfield, H. Dickerman, and N. Brot, JBC 240, 856 (1965). 88. H. F. Kung, C. Spears, and H. Weissbach, ABB 150, 23 (1972).
160
ROBERT T. TAYLOR AND HERBERT WEISSBACH
in the pathway by methionine does not appear to be involved in the cyano-BIZ response. The finding that organisms with low levels of S-adenosyl-L-methionine synthetase have depressed levels of the methionine biosynthetic enzymes (89) also indicates that methionine repression may be mediated by S-adenosylmethionine. These results are summarized below: Methionine
S-adencsylmethionine .--t repreasion of all methionine biosynthetic enzymes
Cyano-Bll + Bt2enzyme -+ specific repreasion of nonBlt transmethylase reductase and, perhaps, N6JO-methylene-H~-folate
IV. Other Sources of Non-B1, and BIZ N5-MethyltetrahydrofolateHomocysteine Methyltmnsfemses
A. NON-B,,METHYLTRANSFERASES 1. Occurrence Except for the methionine-cobalamin auxotrophs, all strains of Eschen'chia coli that have been studied (PA15,K-12,518-W,and B) utilize only the non-B12 enzyme for growth in cobalamin-free minimal media ( l a , 3). In addition to E . coli, the presence of a non-B,, methyltransferase has also been documented for Aerobacter aerogenes (90) and Salmonella typhimurium (91).Both of these enteric non-Blz methyltransferases have substrate and ion requirements that are very similar to the E . coli non-Blz enzyme (90, 91). A key difference between E . coli ( 2 ) and these other two enteric bacteria (90, 9 l ) , however, is that the latter can synthesize enough corrinoids during growth on minimal media to convert partially their BIZrequiring apoenzyme into a B,, holoenayme. Saccuromyces cerevisiae (92, 93) Neurospora crassa (94), Chlorella pyrenoidosa (95), and higher plants ( l a , 36, 53, 96, 97) only contain 89. R. C. Greene, C. H. Su, and C. T. Holloway, BBRC 38, 1120 (1970). 90. J. F. Morningstar and R. L. Kisliuk, J . Gen. Microbbl. 39, 43 (1965). 91. S. E. Cauthen, M. A. Foster, and D. D. Woods, BJ 98, 630 (1966). 92. J. D. Botsford and L. W. Parks, J . Bacteriol. 94, 966 (1967). 93. E. G. Burton and W. Sakami, Fed. Proc., Fed. Amer. SOC.Exp. Biol. 26, 387 (1967). 94. E. Burton, J. Selhub, and W. Sakami, BJ 111, 793 (1969). 95. E. G. Burton, Dissertation, Microfilm No. 71-1667, Univrrsity Microfilms. Ann Arbor, Michigan, 1970. 96. E. G. Burton and W. Sakami, BBRC 36, 228 (1969). 97. W. A. Dodd and E. A. CoaSins, ABB 133, 216 (1969).
4. FOLATE METHYLTRANSFERASES
161
non-B,, methyltransferases. An absence of the B1,enzyme from higher plant cell extracts is consistent with the finding that corrinoids are not metabolites in their tissues except in the root nodules of those species which carry out symbiotic nitrogen fixation (98). Several years ago considerable interest was generated by a report that rat liver also contained a non-Bl, 5-methyl-H,-folate (Glus) transmethylase (99).This claim could not be verified by Burton and Sakami (100). Using [5-’*C] methyl-H4-folate (Glus) (100) instead of unlabeled folate substrate ( 9 9 ) , Burton and Sakami carefully determined that non-B,, transmethylation was insignificant compared to the unlabeled methionine formed by proteolysis and methyl transfer from the endogenous betaine. Therefore, it still appears safe to conclude that mammalian tissues do not contain a non-B12 methyltransferase for folates. 2. Yeast Non-B,, Enzyme Other than from E . coli (Table XIV), a non-B,, folate methyltransferase has been highly purified only from S. cerevisiue (101).The yeast enzyme preparation, estimated to be 85% pure, has a sedimentation coefficient of 5.25 S and a molecular weight of 75,000. It has a catalytic [reaction (2)] specific activity of 3600 mpmoles of methionine formed per milligram per 15 min relative to 2520 mpmoles (Table XIV) for the E . coli enzyme. Yeast non-BI2 enzyme is also specific for L-homocysteine ( K , = 22 as a methyl acceptor, but it utilizes 5-methyl-H4-folate and 5-methyl-H4-folate (Glu,) ( K , = 430 f l ) (Glus) (Kn8= 380 equally well as methyl donors (94). 5-Methyl-H4-folate (GluJ is inactive as a substrate (94, 101). The optimum pH region for the yeast enzyme is 6.6-7.6. It displays an absolute requirement for phosphate, but unlike the E . coli enzyme (Table XV) it is not stimulated by Mg2+ (101). A partial inhibition of its activity by EDTA (96, 101) is suggestive that a bound metal may also be required by this enzyme.
a)
a)
3. Folate Substrates of Non-B,, Enzymes
Until recently it was generally held (7, 94) that the inability to use 5-methyl-H4-folate (Glu,) as a substrate was a distinguishing feature of all non-B,, folate methyltransferases. However, extracts of pea seeds (97), green beans, spinach, and barley sprouts (96) were found subsequently to catalyze 5-methyl-HI-folate (Glu,) -homocysteine trsns98. H. J. Evans and M. Kliewer, Ann. N. Y.Acad. SCi. 11%736 (1964). 99. F. K. Wang, J. Koch, and E. L. R. Stobtad, Biochem. 2. 346, 458 (1966). 100. E. Burton and W. Sakami, Eur. J . Biochem. 7, 1 (1988). 101. E. Burton and W. Sakami, “Methods in Enzymology,” Vol. 17B, p. 388, 1971.
162
ROBERT T. TAYLOR AND HERBERT WEISSBACH
methylation in the presence of Mg2+ and phosphate buffer. Catalysis resulting from the presence of a B,, enzyme in these plant extracts was excluded by the lack of any stimulation with AMe or reduced flavin and the absence of any inhibition by oxygen (96).Upon subjecting extracts of green beans to Sephadex G-100 chromatography, the activities for 5-methyl-H,-folate (Glu,) and (Glus) eluted in the exact same fractions. The monoglutamate substrate was one-seventh as active as the triglutamate (96).Thus, higher plants apparently contain a second type of non-B,, methyltransferase. It resembles the bacterial non-Blz enzyme in requiring only Mg2+and phosphate, but it resembles the bacterial B,, enzyme in utilizing 5-methyl-&-folate (Glu,) as a substrate. Detection of this second type of non-B,, enzyme suggests that the extra L-glutamate groups are required only for binding the folate substrate to the active site of the E. coli (82) and the yeast (94)non-B12 enzymes. It argues against an essential participation of the a-carboxyl group (on the second glutamate residue) in the catalysis per se of reaction (2). At this time, it would seem that the influence of oxygen as opposed to AMe plus a reducing system (3,96) provides the best criteria as to whether one is dealing with a B,, or a non-B12 enzyme in crude systems.
B. B12 METHYLTRANSFERASES 1. Occurrence and Methyl-B,,-Homocysteine Transmethylation
Although B,, methyltransferases have been widely reported in both microorganisms and mammalian tissues, in none of these instances has the eneymc been purified and studied to the same extent as the enzyme from E. coli B (Section 11).I n most cases, evidence for the B12enzyme rests on the presence of reaction (1) activity which shows the unique dependencies on ti reducing system, anaerobiosis, AMe, and a B,, compound if the cells were cultured in the absence of a cobalamin. In con8.I typhimuriurn ), (91),and C . trast to E . coli ( l a , s),A. aerogenes (& pyrenoidosa (96) which contain both types of methyltransferases, only the B,, enzyme has been found in Rhodopseudomonus spheroides (102), Ochromonas malhamensis (103), and Streptomyces olivaceus (104). Similarly, only the B,?enzyme is found in mammalian cells. Thus far, 102. S. E. Cauthen, J. R. Pattison, and J. Lascelles, BJ 102, 774 (1967). 103. J. M. Griffith~and L. J. Daniel, ABB 134, 463 (1969). 104. H. Ohmori, K. Sato, S. Shimisu, and S. Fukui, Agr. Biol. Chem. 35, 338
(1971).
4.
FOLA'IE METHYLTRANSFERASES
163
it has been detected in cell-frcc extracts of pig liver (3, 1.6, 41, 59, 106, 106), beef liver (59), chicken liver (lor),rat liver (108, 109), human liver (110, I l l ) , pig kidney ( I l a ) , human kidney (110, I l l ) , rat brain (113), beef brain (I14),human brain (115),and a variety of mammalian cell lines grown in tissue culture (116-120). Studies of the subcellular distribution of B,* methyltransferase in rat liver (99, 108) have shown it to be located predominantly in the cytoplasmic and mitochondria1 fractions. Brown et al. (108) found 50% in the cytoplasm and 36% in the mitochondria. A survey of the organ distribution of BIZenzyme in bovine tissues showed the pancreas and brain to contain the highest levels of activity (114). Arranged in the order of their enzyme activities per milligram of protein, the following pattern was observed: pancreas > brain > liver > adrenals > heart > kidney. It will be recalled from Section II,C that the.E. coli B,, enzyme catalyzes free methyl-B,,-homocysteine transmethylation [reaction (8) ] in addition to reaction (1). It is noteworthy that in every case where it has been examined, other sources of the B,, enzyme have also displayed this subsidiary activity. This holds true for extracts or partially purified enzyme from R. sphaeroides (102), s. olivaceus (1041, A . aerogenes (53), 8. typhimurium (b3), chicken liver (14, 107),pig liver (14, 36), and rat liver (14). While in none of these systems has the 105. J. H.Mangum and K. G. Scrimgeour, Fed. Proc., Fed. Amer. SOC.Exp. BiOZ.
21, 242 (1962).
106. R. E. Loughlin, H. L. Elford, and J. M. Buchanan, JBC 239, 2888 (1964). 107. H. Dickerman, B. G. Redfield, J. G. Bieri, and H. Weissbach, JBC 239, 2545 (1964). 108. S . S. Brown, G. E. Neal, and D. C. Williams, BJ 97, 34c (1966). 109. C. Kutzbach, E. Galloway, and E.'L. R. Stokstad, Proc. SOC.Exp. Biol. Med. 124, 801 (1967). 110. S. H.Mudd, H. L. Levy, and R. H. Abeles, BBRC 35, 121 (1969). 111. 5. H.Mudd, H. L. Levy, and G. Morrow, Biochem. Med. 4, 193 (1870). 112. J. H.Mangum and J. A. North, Biochemistry 10, 3765 (1971). 113. T. Nakazawa, K. Yoshiba, and M. Takasugi, Bitamin 41, 333 (1970); 42, 193 (1970). 114. J. H.Mangum, B. W. Stewart, and J. A. North, ABB 148, 63 (1972). 115. H.L. Levy, 5. H. Mudd, J. D. Schulman, P. M. Dreyfus, and R. H. Abeles, Amer. J . Med. 48, 390 (1970). 116.J. H.Mangum and J. A. North, BBRC 32, lob (1968). 117.J. H.Mangum, B. K. Murray, and J. A. North, Biochemistry 8, 3496 (1969). 118. 8.H.Mudd, B. W. Uhlendorf, K. R. Hinds, and H. L. Levy, Bwchem. Med. 4, 215 (1970). 119. S. 8. Kerwar, C. Spears, B. McAuslan, and H. Weissbach, ABB 142, 231 (1971). 320. M. J. Mahoney, L. E. Roeenberg, S. H. Mudd, and B. W. Uhlendorf, BBRC 44, 375 (1971).
164
ROBERT T. TAYLOR AND HERBERT WEISSBACH
relationship of reaction (8) activity to reaction (1) activity been studied in as much detail as for the E . coli B enzyme (16),the apparent widespread coexistence of both activities is certainly suggestive that they are properties of a single enzyme. Over a 30-fold enrichment from extracts of chicken liver, both activities copurify together along with the cobalt-60 label from previously injected cyano-60Co-Blz (107). It would be of interest to learn whether any organisms which are reported to contain only the non-B12 methyltransferase also possess any appreciable free methyl-B,,-homocysteine transmethylase activity. 2. Pig Liver and Pig Kidney B,, Enzymes
The most purified and best characterized mammalian enzyme preparations have been those from pig liver and pig kidney. Loughlin et al. (106) purified the enzyme 250-fold from pig liver with an overall yield of 10%. Their preparation showed a nearly complete dependency on reduced flavin and AMe, and its specific catalytic activity [reaction ( l ) ] was 1.0 pmole of methionine synthesized per hour per milligram. The cofactor requirements (106) were, therefore, quite analogous to those of the E . coli enzyme (Table IV). Both the mono- and the triglutamate forms of 5-methyl-H,-folate were equally active as substrates for the pig liver enzyme (106).Most significant, however, was their demonstration (106) that a bound cobalamin is an essential component of a mammalian transmethylase. A constant ratio was observed between the B,, content and the activity content of the fractions a t each step over the 250-fold purification procedure. The final preparations contained about 21 pmoles of cobalamin per milligram of protein (106). Very recently, Mangum and North (119) published in detail a purification scheme to enrich the pig kidney B,, enzyme 1800-fold over crude extracts. The final preparation catalyzed the formation of 54.4 pmoles of methionine per hour per milligram, but it was obtained in an overall yield of only 0.8%. Homocysteine was used to stabilize the protein during its isolation. An absorption spectrum of their best preparation (10 mg/ml) shows (119) that it is probably rather similar in the visible region to the E . coli B enzyme (16). The best pig kidney enzyme is still contaminated by cytochrome absorption in the 410 nm region, however (119). No attempt was made to identify the protein-bound Biz. The only catalytic property reported (119) for 1800-fold purified pig kidney enzyme is. that it has a near absolute dependency on reduced flavin, but is stimulated only 1.3-1.8-fold by AMe. It is suggested by Mangum and North (119) that only a slight dependency on AMe represents a fundamental difference in the mechanism of reaction (1) between
4. FOLATE
METHYLTRANSFWASES
165
the pig kidney enzyme and the E. coli B enzyme (depicted in Fig. 8). However, since the B,, enzyme that was partially purified from pig brain by Mangum et al. (114) showed an absolute requirement for both reduced flavin and AMe, the possibility of nonspecifically bound AMe (49) in the pig kidney enzyme preparation (11a) should be considered. From the combined data in two communications (191, 199),Burke et al. concluded that the mechanism depicted in Fig. 8 essentially accommodates most of the observations made using pig kidney B1, methyltransferase. A key part of the experimental evidence for this conclusion was the accumulation of a [14C]methyl-B,, enzyme from [5-14C]methylH,-folate. The Sephadex G-25 filtered [14C]methyl-B,, enzyme then transferred its [l*C]methyl group to homocysteine (191) under the same conditions as for the E. coli enzyme (Table IX). Folate substrate methylation required AMe, reduced flavin, and 0.15M levels of methionine (131). The high level of methionine required has not been explained. It is possible that high levels of methionine were needed merely to inhibit (122) the transfer of [14C]methyl from the [14C]methyl-B1, protein to the contaminating homocysteine in the enzyme preparation. Partially purified pig kidney methyltransferase can also be methylated with [14C]methyl-AMe,and it catalyzes reaction (6) a t a slow rate compared to reaction (1) (191).
121. G. T. Burke, J. H. Mangum, and J. D. Brodie, Biochemistry 9, 4297 (1970). 122. G. T. Burke, J. H. Mangum, and J. D. Brodie, Biochemistry 10, 3079 (1971).